The Biggest Value Add for Someone Who's Looking to Get into the Space to Make Money: How Do You See the Tool Creation Environment Continuing to Develop and the Importance of It Moving Forward?

I'd like to answer this from a different angle and then I'll come back. A month ago, there were videos saying that the AI AA model is dead, all this kind of stuff. I haven't made money, and I'll tie it back in why this is relevant in a sec.

Think about how many SMBs have websites. Think about how many SMBs you know like it's just you know and all of these SMBs cannot use or it's really difficult to use AI today. And so the way I view it is there's going to be a massive services layer in the middle. Look at how deoe and all these large consultancies are pushing into AI; they're setting up practices or hiring Chief AI officers and all this kind of stuff. It's because if you do not have the skill set to stand stuff up, that's the services layer in the middle, right.

And so how we think about Voice Low is on building the best tools, you're G to have all these SMBs mid-markets out there who do not have internal resources and it does not make sense for them to learn you know how to use Voice Loads. We just use a website contractor because we don't have the time to use uh learn to use web flow to the degree we we need right; it just doesn't make sense to hire a full-time convenience and expertise you'll pay for it every time, totally.

And so I just think about how big the market is going to be for automating things with AI and all the SMBs that will not have the skill set to do it. It's just like okay, it's very clear there's going to be a large services layer in the middle; like that's just like it feels you look at the website market, and I think AI is going to be more impactful than websites. You know because I think we're talking about not only uh the replacement of the interface which we've already kind of talked about but also the replacement of a lot of internal tasks so it's like the RPA market and the interface Market had like this you know this yeah trying to find a good word but yes um.

And so anyways, I'm extremely bullish for that reason of like there's just makes all the sense in the world. So going back to how are good people uh people going to make money? I think people are going to make money setting up agencies and I think when people think about agencies uh you know I mean you've got a ton of great resources on this but like choosing a use case for sure, like building up like building up an expertise and overtime branching out to tenant use cases that's awesome.

Because I think um customers and SMBs don't search for AI automation agency; they search for lead gen automation. You know I mean they search for the use case you just happen to be the model and so I think some people have got that you know kind of backwards thinking that people are searching for the agency name, yeah no one outside the space knows what AI automation the agency is.

No business owner is searching that up totally like we don't you know we're not trying to optimize SEO for like uh you know like collaborative chatbot Builder. We're Auto you know we're doing like customer support automation because people are searching for the use case that the tool solves for so that you know that's sort of the first thing.

So agencies with use cases then I think you're going to have that's all for the video here on the main channel, the rest of it is going to be in its uncut full link format here on the morning side AI channel. So if you want to watch it, you can click up there. If you have enjoyed please hit down below and leave a like and let me know if you want to see more content like this.

And if you haven't already, you can subscribe to the channel for more AI business content.

WEBVTTKind: captionsLanguage: enI saw a whole bunch of videos was like you know the the a AI a model it's dead right all this kind of stuff you know I haven't made money I just think about how big the market is going to be for automating things with AI and all the smbs that will not have the skill set to do it and it's just like okay it's very clear there's going to be a large Services layer in the middle people say oh but they wouldn't pay that for a chat bot it's just a chat bot it's just a this based on conversations like this it becomes clear that we need to reframe it a little bit as to how we look at chat Bo open's goal is not to be a tool Builder they're not making money from the tools they're making money from the consumption and they're making money from having the best models that's a trillion dollar market AI automation's blown up we're just seeing like the proliferation of people realizing that the ROI statement that we've all been used to for the past 10 years has completely flipped and it starts to get you thinking what's the interface even going to look like our websites obsolete now what's going on guys this is an interview between myself and Braden reene the CEO of voice flow I really wanted to get him on the channel to discuss some of the big events and updates within the space regarding the future of chatbot Builders and the future of AI automation agencies in light of these new updates with open AI I had an absolute blast in this conversation with them so I know that there's going to be a ton of value in here for you not only in the direction of the space and where he sees it going but some of the perspective shifts that are required to to properly see the opportunity for what it is I highly recommend you watch the entire video because there's so much good stuff packed into this on the main channel here there's only going to be about half of it with the rest of the the full interview being put on the morning side AI channel so if you want to get to the end of the video and you think you want to see the rest it's going to be linked in the first link in the description we can go over to the morning side AI Channel where I do my proper long form conversations with different people okay welcome everybody to the channel we have a very special guest today Braden ream the CEO of voice flow I've been going back and forth Braden and the team um trying to get this interview because I know there's a lot of stuff going on in the space right now that people are dying to hear my thoughts on but more importantly I thought I get it from the horse's mouth um as someone who's leading a company in the space and something that myself and and my community use heavily is to get Bren on here and give his thoughts on what happened with the assistance API the gbts update and how this affects the space moving forward so Braden firstly welcome um it's it's a pleasure to have you on the channel and uh looking forward to hearing what you have to say yeah excited to be here uh it's been a crazy couple of months and so it be a lot of fun to chat about all that today awesome well we just jump straight into it and go roll back what two two and a bit weeks now um the gpts r update rolled out the D Day announcements um all hell broke loose it seemed um what's your take on the recent updates and I think specifically if we can start with the the trend towards uh autonomous agents and instead of human input we're going seem to AI is leading us a little bit more towards a future where it's not necessarily human input but more uh sort of agent to agent communication where do you think uh voice flow fits into this kind of industry and space moving forward totally so I think you know before I even talk about voice flow um I'll talk a little bit about sort of the history of assistance uh because I think it it plays into this nicely so um when we got into this space in like 2018 2019 it was Alexa that was all the rage you know that was sort of the big dominant consumer assistant and I don't think a lot of people paid attention to how Alexa was trying to build their ecosystem but what they were trying to build is Alexa skills and Alexa skills were essentially um thirdparty apps you could build on top of Alexa and what was really interesting is Alexa in the same way that uh opening ey gpts now is trying to uh connect these different thirdparty apps had the same thing skills connections and they really laid out the model for having a consumer assistant that would be like an Alexa that is essentially a browser for the internet of assistant out there the internet of agents and what's interesting is over time uh you know you would interact with your consumer assistant who could then access you know the business assistant right so maybe I wanted to uh book a hotel with Expedia or something like that today you would go through Alexa and again I'm using Alexa but you could you know swap in chat gbt or really any assistant you could talk to ch gbt then talk to Alexa and book a hotel or sorry talk to Expedia book a hotel what really starts to become interesting is that's just one task like let's say you want to go to Paris or something you need a hotel you need flights you need like a bunch of different Services when can these agents and assistants start to do compound tasks when they can start to decide what do I need in order to actually fulfill this entire this entire journey and I think that's where um Alexa skills was trying to go and that's now where sort of the successor with uh gpts is is now really heading in that direction and even plugins earlier uh this year you know it looked like it was going to be it right yeah totally I mean you know we had this we had a lot of this similar conversations when plugins came out um earlier this year that we're now having about gbts and it's just really cool because we're now moving towards um this sort of Internet of agents and I think you're going to see essentially every business will have an agent and every uh consumer will have a consumer assistant that they trust right and so I think over time you're going to start to see um people are going to choose you know maybe they use Alexa as their shopping Focus agent maybe it's Siri which is going to be their privacy focused agent maybe it's going to be chat gbt which is you know the all-encompassing super smart agent and it' be really interesting because you sort have these consumer agents that then interface with the business agents so I think that's you know if we go the at the highest level that's sort of the macro um when it comes to uh some of the announcements that came out with uh the gpts my opinion is that um open AI is trying to build the best models they're trying to get the AGI and in order to do that you need um the best data and so what they've done is they they're trying to turn a model advantage that uh I think this year what we've seen is like the model advantages are more fleeting than we initially thought like companies are catching up within months right and so um they're launching these models and then now they're trying to turn that short-lived model Advantage into like a distribution advantage that can then lead to better data so they launched uh gpts which allows you to do these like specialized use cases on top which then allows them to get uh a distribution Moe because now like chat gbt is becoming useful not just for the model but for the application layer on top of it um and also it's giving them a way better Specialized data because now it's not just chat gbt that's like very broad you know the conversation's not labeled it's very specifically uh you know a lead gen uh GPT it's very specifically an accounting GPT so that data is better than for them to be able to use because it's it's more filtered right the better quality data the better quality models you get and so the way that we view it is open ai's goal is to get better data to eventually get the AGI to build you know to build better models and this is just a great way to create a distribution mode to make sure that they continue to get more data that chat gbt continues to be the place that we all interact with and um that the gpts give them uh higher quality more sort ofine tune data now how do you do that imagine you're opening eye for a second you have to create tools that make it easy for people to actually Kickstart that process um and I don't know if we want to go too deep here but every we can go deep yeah yeah so every product follows this like fundamental law which is you can either be easy or powerful and the reason why that is is because every feature you add like you know for all the chatbot Builders uh sort of watching every time you add an additional widget or button or you know some sort of capability it makes the chatbot more powerful but it also makes it a little bit harder to use right the easiest chatbot would be something that is um so you know just button- driven right really simple two or three flows um you know it doesn't do much but it's bulletproof um same thing with product design so we think about this at voice flow a lot every button we add gets more more and more powerful right every feature every whatever whatever uh we add the problem is as you do that it becomes way more complex and hard to use and so like let's say website builders for example do you want to be Wix or do you want to be webflow or WordPress they all fall in this like spectrum of easy to use to hard to use hard to use typically skews to more value um it's a more powerful platform typically like mid-market uh Enterprise really you know uh skilled agencies who then service those larger customers or skilled agencies who Service uh maybe an SB based that wants more than just like a Wix website so when we think about what open AI wants to do they want data right to build the best models so they're going to build tools that enable the widest audience of people which means the tools are going to be pretty simple um I think over time as they need to uh get better data they might make the tools more complex but they risk if they make the tools too complex they risk sort of having the broadest top of funnel right because if you make a a builder so easy that everyone can build a TPT yes the build might not be very happy but opening ey's goal is not to be a uh a tool Builder they're not making money from the tools they're making money from the consumption and they're making money from having the best models that's a trillion dollar market they don't care about the tool Market which is much smaller and so the tools are just there as a way to facilitate um the outcome But ultimately I think the like sort of the power and complexity they're going to leave to third party tools like a voice flow for example um because we're just going after a smaller market in terms of the number of people out there higher value you know more complex and things like that but their Market way bigger and they're just trying to get the data so um that's kind of how we think about it you know it comes at it from a couple different lenses there um hope I'm glad you said that because I've literally it's it's so good to hear it from another person especially someone in your position because am I someone in my comment section and said uh like what's the I don't know there was a couple different ones one was um why wouldn't what's the point of building on these platforms anymore wouldn't open AI just roll out the own feature for themselves and aren't they one step away from creating a chat Builder and all this kind of stuff and my take was was exactly that which is that's not the that's the pennies game they're in the the big money is in the foundational models and in the building the the API that everything runs on for them if if you think put yourself in the position of samman or whoever now samman again but would you be bending down to pick up the relative pennies in the market which are still massive markets but relative to the the foundational models side of things it's just not if you had limited resources you wouldn't be allocating it to oh but there's a little bit of money there make a Instagram WhatsApp and telegram deployer straight out of a like it just doesn't make sense for them they can focus on what they're good at rather than getting their hands Tangled into and I'm sure you know as soon as you get into uh interfaces and things like this it's just a completely different ball game so if they could stick to their their back end work their their apis I'm sure that's where they' they'd much rather be sitting I mean they're A90 billion company vo is a100 million doll company um and like you know they're growing like bananas um and you know if we actually take this down to the the tangible voice flow is not easy to use right like relative to uh how easy it could be right you know you come in you know we could we could strip out a ton of the complexity but then you lose a lot of that power so if they were to build what voice flow has they can for for example they could do some of the business logic stuff for sure and some of the Integrations um but then you start to alienate the audience and then what's interesting is then other third party tools that make it easier will pop up right and so they have to think about do they want to be the biggest widest most Mass Market or do they want to be uh the biggest most complex and I think um so boy flow has what anthropic models open a ey models and we got other models uh coming um they're still getting data and consumption from supporting third party tools and so we've got a great partnership with them where or you know maybe not a formal partnership we got a good relationship I should say um and we're still bringing them great big Enterprises who are now using vo those more complex tools and feeding that data through to open AI so they kind of win both markets by uh exposing these apis but then enabling people who can't use the more complex more powerful tools with really easy simple tools they actually kind of capture both sides of the market yeah and I think the the interface is a interface side of the whole equation is a very interesting one and it's one that you guys have made a bit of a pivot recently on away from necessarily having WhatsApp and these different deployments down to just the voice and the chat deployments essentially so what is your thoughts on on the future of interfaces and this this big question mark that is in the in the sort of software 3.0 stack as I believe what some people call it where do you see the interface going uh for for AI based applications and how is voice flow uh tailoring its its strategy if you want to go into strategy but how we can expect you guys to play this moving forward totally I mean there's so much Innovation like I think the interfaces is um there's going to be so much change over the next year on the interface side that our approach was actually to pull back and just focus on what we're really really good at which is the building experience and then expose an API and allow others to build their own interface so we have our web chat widget which is essentially the one channel that we're going to uh natively support really really well and then everything else is going to be done through an API and the reason for this is we had customers com and say hey I'm doing an automated drive-thru and I need like a face right can you guys build an interface for that uh I'm doing a SMS integration with this provider I'm doing uh WhatsApp I'm doing like you know in I'm doing all these different interfaces especially the custom interfaces there's a lot of those and we started to realize that um we're not going to be the company to build out those interfaces you know we've got 200,000 users now like how do we enable our users to be able to build interfaces for each other to make it really easy to build their own interfaces customize these interfaces because I think um you know a really interesting Trend I've seen is you go to a website and you know when the web chat widget first launched as sort of like a paradigm I don't know like five six seven years ago um the Bots were not very powerful ful and so you would you know use the website then you use the widget for a little bit then you go back to the website the Bots are now getting so good that we've seen where like you click on the you know you click on the widget and it just goes full screen because there's no point in even using the website anymore and it starts to get you thinking like what what's the interface even going to look like like is web are websites obsolete now um as a paradigm because like I'm not even using the website I'm just using the chatbot and so you start to think well hold up like if the chatbot can start to give you know more detailed responses will when you land on a website page in future is it just going to be a chat bot and it will serve you the website elements as needed right and as you have a conversation with this website and so I think like that kind of interface uh Innovation is going to be so fast and Rapid that like we're going to let the community figure out how they want to do that and then just try to support it as best we can um so that's sort of the approach right now and then in future you can expect us to add more what we would call like direct Channel Integrations again so whether that's Instagram or Whatsapp but we just really want to focus on like enabling the community to innovate right now um and then over time I think we'll start to add more and more like native Channel Integrations it's funny you mention that as well because I was having a chat with my uh my template team on the the accelerator and had the same conversation I was just like racking my brain like where does this where does the interface go like how does this where does where does this all lead and what ises the interface of the future look like and how we interact with these assistants and Abby one of our one of our developers in there who was saying when like I I think the the interface becomes becomes chat in many ways and it does like you were saying make make the website in in many aspects obsolete because with the the ability to call functions throughout it as we saw with the DFD demo is that being able to converse with the assistant and be able to do things that you normally have to do by clicking around it's as I've said in Prior videos it's the end of this this point and click websites and it's trans sort of transforming over to a a conversational based interface but I think it's also important and a good an important reframe to make for people who are working in the space is that chat bots in the way that we know them have a have a basic and and sort of mundane and dull sound to them in many ways but yet we have it is literally the interface of the future at least before we get to the voice side of things I think it's been difficult pill for some people to swallow um it's something that I've been kind of having to push our pill along my journey is that people say oh but they wouldn't pay that for a chat bot it's just a chat bot it's just a this and based on conversations like this becomes clear that we need to reframe it a little bit as to how we look at chatbots it's not just a chatbot it's a it's a the new new way of interacting with technology in any ways so I'm glad to hear you you've got the same thoughts on that it is very exciting the thought of we've even considered to put it on our um on my on my accelerator website where it's okay you come on instead of having to read through the whole page people have questions that come to their mind as soon as they get to the page have to make them pick through the entire website to try to find it why not just ask the question and we can we can give you the answer and we can serve you up resources and links and everything that you need and I think the assistance API has been a big step towards that with the knowledge based component and the tools so I love the idea of reframing what a chat bot is um you know this is something when we were raising money for voice flow we had the same thing from investors you know who cares about chat Bots and you know the way that we frame this to folks is you know when you think about a chat bot you just think about like the chat bubble but it's so much more than that if you if you roll it back and you start to think about every single interaction that you have with a customer is a part of a conversation so when I'm on Netflix and I'm trying to choose a movie to watch and you know it's I've been scrolling for two minutes now that's me like you know in the most abstract way saying hey hey Netflix I can't find what I want can you jump in here and suggest something right like that was an implicit act in a conversation Netflix then jumps in over time you know uh I need to add another subscription I need to you know chat with them over text like all the different channels all the different interfaces all the different both implicit and explicit actions s to blend to uh chatbots essentially being like the CRM like the relationship operating system between uh the customer and your business and yeah and and like I think when people start to frame it that way that it is um the chat bubbles is the interface that people think about what people think that the interface is the product but the product is the agent the product is like that behind the scenes tracking all different interactions with all your customers and done really well it can be just an awesome experience uh like what if ever had it where you know you call an ivr and uh like a you know like a contact center and it already knows what you know it already knows what you're calling about you know because it looked up your number it saw you know who you are as a customer it saw that you purchased something last week like when you have experiences like that where again it feels like wow like this is not just some call center bot there's some intelligence behind the scenes like they have an agent there um that's when like the experiences get really magical people need to start thinking about disambiguating the interface that you see and the behind the scenes logic that is the operating system between how your business interacts with customers for sure I'm I'm very big on this this data side of things I think the the conversational a conversation as a channel via whatever the you want to use there's something special about being able to to pull information out of those conversations because you get to see into the customers world and life in a way that you don't really get to when they're just interacting with your website because they don't think that those conversations are being mon like it's not top of mind when you're asking something to a web chat widget about a product or about you're not thinking that this is going to be dug over and analyzed over time for for information and especially when it comes down to those more personal ones where it's not so much just customer support but it's actually like how can I help and it's getting very very personal with that and feedback as well there's there's so much to be minded out of this and I was looking up a good conversational data you might know something something like this but a like industrial scale or like business scale conversational data analytics that uh you can plug into something like these these assistants now as you're pumping out all these threads there needs to be a way I've just done a video on the channel um covering how you can export these threads and then analyze them um using air table to make it's nothing nothing too fancy but there's going to be massive need for okay I've got all these threads that have just been created over the past week where can I plug them in and get some insights out of it so do you know if there's any conversational data analytics platforms out there or or what do you guys use internally for that if you're allowed to say yeah I mean we um we've got some features and boic that aren't publicly available that do stuff like this and I'll maybe I'll show you after the recording um but I think a lot of people when they think about chat Bots again they're thinking on like the very first level which is like I see the chat Bubbles and I see it just for one particular use case like let's take lead gen but um I think what gets really interesting and I think this is for a lot of the um agency owners the very first step when you start working with a client and we're trying to do this more rigorously ourselves as well is let's analyze the data that you have let's analyze the transcript data like where is the opportunity for automation because you might think it's leg gen but it might be some internal uh task automation right and so for example you know we were chatting with a company the other day where um they had a bunch of uh I think they have like 600 cold call uh agents uh and what they need was uh lead enrichment and that was it like they just needed to be able to say hey I'm going to call this person run it through a chat bot that says hey I just pulled up older data here's like the best way to approach this cold call script M that by 600 and then like people 50 calls a day like 100 calls a day totally factoring the time of day factoring any potential holidays like and you start to realize that for that it's a very bespoke use case right leave lead jet is what I would consider um a very high volume high impact use case but where there's a lot of opportunity is um there's a thousand small little use cases at most businesses like you know we have use cases that only exist at Voice Low right things that you know only we uniquely need and how can we um work with companies that are automating not just like the core pillars everyone knows customer support everyone knows customer experience how can we start automating a lot of those bespoke um like mid mid volume low volume use cases that have high impact right so like we're starting to think a lot about this now is voo not just being a way to um you know automate customer support but it's like customer support's just the beginning once you're in you know once you've landed with a customer how do you then automate everything else right and really thing about like that AI automation piece which is why like I love sort of where um your s and others like are taking the space um because that's totally what it is it's AI automation but everyone needs everyone's focusing on the same use cases there's so many long use cases that provide a ton of value um and I think folks aren't necessarily thinking about it that way yeah I think that that comes down to me quite a bit I think I do try to at least in most of my content to make it as beginner friendly as possible and it does in order to attack those Niche use cases and the the business specific ones that gets a bit closer to a custom solution and in my experience you need a a fairly deep and and and Broad skill set as well in order to be able to piece together a solution for that so that includes like the automation platforms the chat Builders the the backend sort of more complex knowledge based functionality that you need so the I feel like the pillars are a good point for people to get into like a customer support leion and all these kind of areas where it's easy and and they understand it and there's places you can get templates like I've put lead generation templates and these kind of things out that are fairly easily attackable but that is the definitely the destination that we want to get a lot of the agency owners in the community to which is being able to come in there and and take a full sort of head to toe look at the business and identify those combination of of high impact and and low cost to deliver as well so it's it's ideally it's not ridiculously hard for you to build at the same time which is that sweet spot between valuable yet not too hard to deliver totally right like it's um and we think about this the exact same way um we have pillar use cases like you know our sales and marketing team very focused on customer support customer experience Automation in app co-pilots like you know sort of like the pillar use cases that people know about but what we're starting to do with uh our larger customers to say hey um if you're comfortable with it give us your call transcript data and let's see um you know you can do methods like utterance clustering let's look at the 10,000 utterances and let's cluster them into groups and you might realize hey there's actually this whole group of automation that's not even being touched right let's start looking at internal opportunities internal conversations um like conversational uh Analytics it's going to be a really big area for sure like um now that you can kind of automate everything like if you look at the history of chatbots over the past like you know 10 years in so of these know kind of came onto the scene in 2014 um chots were so expensive to build before llms that customer support was the only use case that even made sense from an Roi standpoint that was it it was like big companies had maybe their own custom teams like we work with a lot of big companies who you know they spend you know there you can go find some reports online $100 million building their chat bot is was a very big Bank in America that was how much they spent send my way next time they come around can deliver something there discount it's it's crazy so you see like a price tag like that I can tell you like from working with some of these teams there's hundreds of people like hundreds building these these Bots and a lot of them were built on the nlu Paradigm where that was what was required now with large language models you can have a two to three person agency stand something up that's like pretty comparable um and that's what's now leading to this explosion of like smbs uh you know midm markets being like hey like I don't need to have a $100 million budget to put together an assistant I can now do this with a very small team and that's what's been really exciting because it's now unlocking all these other use cases as well where again customer support was the only thing that had an Roi that made sense now it's like hold up if I can actually enable that you know 600 person call group you know and it's not going to cost me a million dollars to put that bot together that Roi is now actually worth it I'm going to go do it um and so I think we're just seeing like and I think it's like um the AI automation blown up we're just seeing like the proliferation of people realizing that the ROI statement that we've all been used to for the past 10 years has completely flipped everything is now for the most part like Roi positive as long as you have the right tools and skill set sure I think it'd be good to to also discuss what you've see to be as the the tech stack moving forward and how voice fits into that and yes we have the assistance API which is going to be running a lot of things we have some token costs that I want to talk to you about in a second about and the sort of usability of that in its current state but there's things like the assistance API there's voice flow and there's obviously the the major component of this which is tooling how do you see that sort of ecosystem tying together and how people can play into that and and build sort of solutions uh using your software yeah so how we think about voice flow is largely as an orchestration uh framework and so um how do you orchestrate multiple different models multiple different gpts um we're of the belief that uh if you want to build complex multi-turn um agents that are you know um reliably able to automate not just knowledge but tasks right because we sort of at voice how the internal framework we use is there's knowledge automation that's like your knowledge based Q&A like that's sort of that kind of stuff and then um there's the task automation how do we actually do refunds you know uh TI you know collect emails like that kind of stuff integration yeah essentially integration heavy stuff um we're very focused on okay cool you have you're gonna have gpts which are going to do a lot of the knowledge uh automation I think over time they'll get into some of the simple Integrations do we think open AI is going to be the only player in the space no so how do we be like the model agnostic orchestration framework that's really collaborative really powerful can scale all the way up from like you know a really uh Advanced agency all the way to the world's largest companies you know we have a bunch of features coming for scale soon um and that's how we're thinking about ourselves uh agnostic uh all theit orchestrating the different components because we don't think that for example let's say someone Builds an awesome GPT to do uh you know function a that's great have that like orchestrate with voice flow we've been showing off a lot of those types of capabilities recently and even as the gpts start to talk to each other we think that the ecosystem is going to be large enough where it'll never just be fully contained with open Ai and if it is that's totally cool at the end of the day we're just going to help customers build the best experiences but we believe in a world that's going to be a multimodel so there's gon to be a lot of different models out there you're going to want to keep uh some of your Integrations not just entirely with a no open AI you might want to have a third-party Integrations framework which is actually uh maybe you know hosted on a private cloud or whatever it might be um that's kind of how we're thinking about voice flow from from the long run um scalability extensibility um and then ultimately just trying to be agnostic to all the underlying models yeah I think we're on the same page there with just started work on our I'll probably have chat to you and Daniel about it a bit more and to get your thoughts on it but we are sort of rolling out our model agnostic platform that's going to help people to a lot of the agency owners to hook into platforms like voice flow and so it's sort of a little intermediary step between open Ai and and voice flow that's going to allow people to connect into it very easily and I'm super excited to get more information on about that but the uh I think we've seen with this open AI episode which turned our lives upside down I'm sure well it did for me I can only imagine for you with with the company you have on your hands um how much this past the past week must have felt but the the need for diversification away from why model provider has been really highlighted here for the past couple of days until Sam mman announced the the return I was thinking okay does this mean we've got if they're they're reallocating all their resources to be a research company again this means the assistance API is not going to get the development of needs it's it's still too slow it's still too expensive in many ways they haven't done memory management ET there's so much that needs to be done there but it was so promising it was really concerning like are we going to have to set and and TT thumbs here for 6 months until another company picks up the slack but thankfully we've now got them back on the scene however however that's that's going to play out but anord coming in with CLA 2.1 with tools I think is a very good sign for for what is to come and allowing us to be not so riant on just one provider so you said you have allowed and clawed support recently yeah yes yes we've got anthropic support uh all the open AI models actually I think I think we haven't launched turbo just yet but uh for gbd4 um yeah and anthropic um open a and then uh we've got plans to add a lot more models as well as custom models as well because we you know we work with some big Enterprises where they're just building their own stuff their own large language models uh and I think also with all the open source large language models uh coming out now it's I I think it's more likely than not that you have open AI remains kind of like the dominant close Source player and then you're gonna have just a slew especially on the big company side of like open source models um because it allows you control your data uh you know you don't have to worry about Cloud outage and things it's all in your Cloud it'll be really interesting to see how how much it fragments MH yeah I think tooling is going to be for for agency owners and people looking to make money in the space is to sort of ride the back of the stuff that's coming out what is your thoughts on uh tooling and the importance of knowing how to build their moving on because I feel like that out of the what I call like three ingredients out of these gbts and assistance you prompting knowledge and tooling any any idiot can do the the first two but being able to create custom tools that are powerful and specialized to a certain use case or adapted to a business I see is as what's going to be the biggest value add for someone who's looking to get into the space to make money how do you see the uh the tool creation environment continuing to develop and and the importance of it moving forward totally I actually want to answer this from a different angle and then I'll come back what I want to address is you know maybe a month ago there was uh I saw a whole bunch of videos it was like you know the the a AI AA model it's dead right all this kind of stuff you know I haven't made money um and I'll tie it back in why this is relevant in a sec but I thought it was such a like a ridiculous thing for this reason of think about how many like smbs have websites think about how many smbs um you know like it's it's you know and all of these smbs cannot use or it's really difficult to use AI uh today and so the way I view it is there's going to be a a massive Services layer in the middle I mean look at how deoe and all these like large consultancies are pushing into AI they're setting up practices or hiring You Know Chief AI officers and all this kind of stuff it's because if you do not have the skill set to stand stuff up that's the services layer in the middle right and so how we think about is Voice Low focus on building the best tools you're G to have all these smbs midm markets out there who do not have internal resources and it does not make sense for them to learn you know how to use voice loads we you know we use a website contractor because we don't have the time to use uh learn to use web flow to the degree we we need right it just doesn't make sense to hire a fulltime convenience and expertise you'll pay for it every time totally and so um I just think about how big the market is going to be for automating things with AI and all the smbs that will not have the skill set to do it and it's just like okay it's very clear there's going to be a large Services layer in the middle like that's just like it feels you look at the website market and I think AI is going to be wor impactful than websites you know because I think we're talking about not only uh the replacement of the interface which we've already kind of talked about but also the replacement of a lot of internal task so it's like the RPA market and the interface Market had like this you know this this yeah trying to find a good word but yes um and so anyways I I'm extremely bullish for that reason of like there's it just makes all the sense in the world so going back to how are good people uh people going to make money I think people are going to make money setting up um a gencies and I think when people think about agencies uh you know I mean you've got a ton of great resources on this but like choosing a use case for sure like building up like building up an expertise and overtime branching out to tental use cases that's awesome because I think um customers and smbs don't search for AI automation agency they search for lead gen automation you know I mean they search for the use case you just happen to be the model and so I think some people have got that you know kind of backwards thinking that people are searching for the agency name yeah no one outside the space knows what AI automation the agency is no no business owner is searching that up totally like we don't you know we're not trying to optimize SEO for like uh you know like collaborative chatbot Builder we're Auto you know we're doing like customer support automation because people are searching for the use case that the tool solves for so that you know that's sort of the first thing so agencies with use cases then I think you're G to have that's all for the video here on the main channel the rest of it is going to be in its uncut full link format here on the morning side AI channel so if you want to watch it you can click up there if you have enjoyed please hit down below and leave a like and let me know if you want to see more content like this and if you haven't already you can subscribe to the channel for more AI business contentI saw a whole bunch of videos was like you know the the a AI a model it's dead right all this kind of stuff you know I haven't made money I just think about how big the market is going to be for automating things with AI and all the smbs that will not have the skill set to do it and it's just like okay it's very clear there's going to be a large Services layer in the middle people say oh but they wouldn't pay that for a chat bot it's just a chat bot it's just a this based on conversations like this it becomes clear that we need to reframe it a little bit as to how we look at chat Bo open's goal is not to be a tool Builder they're not making money from the tools they're making money from the consumption and they're making money from having the best models that's a trillion dollar market AI automation's blown up we're just seeing like the proliferation of people realizing that the ROI statement that we've all been used to for the past 10 years has completely flipped and it starts to get you thinking what's the interface even going to look like our websites obsolete now what's going on guys this is an interview between myself and Braden reene the CEO of voice flow I really wanted to get him on the channel to discuss some of the big events and updates within the space regarding the future of chatbot Builders and the future of AI automation agencies in light of these new updates with open AI I had an absolute blast in this conversation with them so I know that there's going to be a ton of value in here for you not only in the direction of the space and where he sees it going but some of the perspective shifts that are required to to properly see the opportunity for what it is I highly recommend you watch the entire video because there's so much good stuff packed into this on the main channel here there's only going to be about half of it with the rest of the the full interview being put on the morning side AI channel so if you want to get to the end of the video and you think you want to see the rest it's going to be linked in the first link in the description we can go over to the morning side AI Channel where I do my proper long form conversations with different people okay welcome everybody to the channel we have a very special guest today Braden ream the CEO of voice flow I've been going back and forth Braden and the team um trying to get this interview because I know there's a lot of stuff going on in the space right now that people are dying to hear my thoughts on but more importantly I thought I get it from the horse's mouth um as someone who's leading a company in the space and something that myself and and my community use heavily is to get Bren on here and give his thoughts on what happened with the assistance API the gbts update and how this affects the space moving forward so Braden firstly welcome um it's it's a pleasure to have you on the channel and uh looking forward to hearing what you have to say yeah excited to be here uh it's been a crazy couple of months and so it be a lot of fun to chat about all that today awesome well we just jump straight into it and go roll back what two two and a bit weeks now um the gpts r update rolled out the D Day announcements um all hell broke loose it seemed um what's your take on the recent updates and I think specifically if we can start with the the trend towards uh autonomous agents and instead of human input we're going seem to AI is leading us a little bit more towards a future where it's not necessarily human input but more uh sort of agent to agent communication where do you think uh voice flow fits into this kind of industry and space moving forward totally so I think you know before I even talk about voice flow um I'll talk a little bit about sort of the history of assistance uh because I think it it plays into this nicely so um when we got into this space in like 2018 2019 it was Alexa that was all the rage you know that was sort of the big dominant consumer assistant and I don't think a lot of people paid attention to how Alexa was trying to build their ecosystem but what they were trying to build is Alexa skills and Alexa skills were essentially um thirdparty apps you could build on top of Alexa and what was really interesting is Alexa in the same way that uh opening ey gpts now is trying to uh connect these different thirdparty apps had the same thing skills connections and they really laid out the model for having a consumer assistant that would be like an Alexa that is essentially a browser for the internet of assistant out there the internet of agents and what's interesting is over time uh you know you would interact with your consumer assistant who could then access you know the business assistant right so maybe I wanted to uh book a hotel with Expedia or something like that today you would go through Alexa and again I'm using Alexa but you could you know swap in chat gbt or really any assistant you could talk to ch gbt then talk to Alexa and book a hotel or sorry talk to Expedia book a hotel what really starts to become interesting is that's just one task like let's say you want to go to Paris or something you need a hotel you need flights you need like a bunch of different Services when can these agents and assistants start to do compound tasks when they can start to decide what do I need in order to actually fulfill this entire this entire journey and I think that's where um Alexa skills was trying to go and that's now where sort of the successor with uh gpts is is now really heading in that direction and even plugins earlier uh this year you know it looked like it was going to be it right yeah totally I mean you know we had this we had a lot of this similar conversations when plugins came out um earlier this year that we're now having about gbts and it's just really cool because we're now moving towards um this sort of Internet of agents and I think you're going to see essentially every business will have an agent and every uh consumer will have a consumer assistant that they trust right and so I think over time you're going to start to see um people are going to choose you know maybe they use Alexa as their shopping Focus agent maybe it's Siri which is going to be their privacy focused agent maybe it's going to be chat gbt which is you know the all-encompassing super smart agent and it' be really interesting because you sort have these consumer agents that then interface with the business agents so I think that's you know if we go the at the highest level that's sort of the macro um when it comes to uh some of the announcements that came out with uh the gpts my opinion is that um open AI is trying to build the best models they're trying to get the AGI and in order to do that you need um the best data and so what they've done is they they're trying to turn a model advantage that uh I think this year what we've seen is like the model advantages are more fleeting than we initially thought like companies are catching up within months right and so um they're launching these models and then now they're trying to turn that short-lived model Advantage into like a distribution advantage that can then lead to better data so they launched uh gpts which allows you to do these like specialized use cases on top which then allows them to get uh a distribution Moe because now like chat gbt is becoming useful not just for the model but for the application layer on top of it um and also it's giving them a way better Specialized data because now it's not just chat gbt that's like very broad you know the conversation's not labeled it's very specifically uh you know a lead gen uh GPT it's very specifically an accounting GPT so that data is better than for them to be able to use because it's it's more filtered right the better quality data the better quality models you get and so the way that we view it is open ai's goal is to get better data to eventually get the AGI to build you know to build better models and this is just a great way to create a distribution mode to make sure that they continue to get more data that chat gbt continues to be the place that we all interact with and um that the gpts give them uh higher quality more sort ofine tune data now how do you do that imagine you're opening eye for a second you have to create tools that make it easy for people to actually Kickstart that process um and I don't know if we want to go too deep here but every we can go deep yeah yeah so every product follows this like fundamental law which is you can either be easy or powerful and the reason why that is is because every feature you add like you know for all the chatbot Builders uh sort of watching every time you add an additional widget or button or you know some sort of capability it makes the chatbot more powerful but it also makes it a little bit harder to use right the easiest chatbot would be something that is um so you know just button- driven right really simple two or three flows um you know it doesn't do much but it's bulletproof um same thing with product design so we think about this at voice flow a lot every button we add gets more more and more powerful right every feature every whatever whatever uh we add the problem is as you do that it becomes way more complex and hard to use and so like let's say website builders for example do you want to be Wix or do you want to be webflow or WordPress they all fall in this like spectrum of easy to use to hard to use hard to use typically skews to more value um it's a more powerful platform typically like mid-market uh Enterprise really you know uh skilled agencies who then service those larger customers or skilled agencies who Service uh maybe an SB based that wants more than just like a Wix website so when we think about what open AI wants to do they want data right to build the best models so they're going to build tools that enable the widest audience of people which means the tools are going to be pretty simple um I think over time as they need to uh get better data they might make the tools more complex but they risk if they make the tools too complex they risk sort of having the broadest top of funnel right because if you make a a builder so easy that everyone can build a TPT yes the build might not be very happy but opening ey's goal is not to be a uh a tool Builder they're not making money from the tools they're making money from the consumption and they're making money from having the best models that's a trillion dollar market they don't care about the tool Market which is much smaller and so the tools are just there as a way to facilitate um the outcome But ultimately I think the like sort of the power and complexity they're going to leave to third party tools like a voice flow for example um because we're just going after a smaller market in terms of the number of people out there higher value you know more complex and things like that but their Market way bigger and they're just trying to get the data so um that's kind of how we think about it you know it comes at it from a couple different lenses there um hope I'm glad you said that because I've literally it's it's so good to hear it from another person especially someone in your position because am I someone in my comment section and said uh like what's the I don't know there was a couple different ones one was um why wouldn't what's the point of building on these platforms anymore wouldn't open AI just roll out the own feature for themselves and aren't they one step away from creating a chat Builder and all this kind of stuff and my take was was exactly that which is that's not the that's the pennies game they're in the the big money is in the foundational models and in the building the the API that everything runs on for them if if you think put yourself in the position of samman or whoever now samman again but would you be bending down to pick up the relative pennies in the market which are still massive markets but relative to the the foundational models side of things it's just not if you had limited resources you wouldn't be allocating it to oh but there's a little bit of money there make a Instagram WhatsApp and telegram deployer straight out of a like it just doesn't make sense for them they can focus on what they're good at rather than getting their hands Tangled into and I'm sure you know as soon as you get into uh interfaces and things like this it's just a completely different ball game so if they could stick to their their back end work their their apis I'm sure that's where they' they'd much rather be sitting I mean they're A90 billion company vo is a100 million doll company um and like you know they're growing like bananas um and you know if we actually take this down to the the tangible voice flow is not easy to use right like relative to uh how easy it could be right you know you come in you know we could we could strip out a ton of the complexity but then you lose a lot of that power so if they were to build what voice flow has they can for for example they could do some of the business logic stuff for sure and some of the Integrations um but then you start to alienate the audience and then what's interesting is then other third party tools that make it easier will pop up right and so they have to think about do they want to be the biggest widest most Mass Market or do they want to be uh the biggest most complex and I think um so boy flow has what anthropic models open a ey models and we got other models uh coming um they're still getting data and consumption from supporting third party tools and so we've got a great partnership with them where or you know maybe not a formal partnership we got a good relationship I should say um and we're still bringing them great big Enterprises who are now using vo those more complex tools and feeding that data through to open AI so they kind of win both markets by uh exposing these apis but then enabling people who can't use the more complex more powerful tools with really easy simple tools they actually kind of capture both sides of the market yeah and I think the the interface is a interface side of the whole equation is a very interesting one and it's one that you guys have made a bit of a pivot recently on away from necessarily having WhatsApp and these different deployments down to just the voice and the chat deployments essentially so what is your thoughts on on the future of interfaces and this this big question mark that is in the in the sort of software 3.0 stack as I believe what some people call it where do you see the interface going uh for for AI based applications and how is voice flow uh tailoring its its strategy if you want to go into strategy but how we can expect you guys to play this moving forward totally I mean there's so much Innovation like I think the interfaces is um there's going to be so much change over the next year on the interface side that our approach was actually to pull back and just focus on what we're really really good at which is the building experience and then expose an API and allow others to build their own interface so we have our web chat widget which is essentially the one channel that we're going to uh natively support really really well and then everything else is going to be done through an API and the reason for this is we had customers com and say hey I'm doing an automated drive-thru and I need like a face right can you guys build an interface for that uh I'm doing a SMS integration with this provider I'm doing uh WhatsApp I'm doing like you know in I'm doing all these different interfaces especially the custom interfaces there's a lot of those and we started to realize that um we're not going to be the company to build out those interfaces you know we've got 200,000 users now like how do we enable our users to be able to build interfaces for each other to make it really easy to build their own interfaces customize these interfaces because I think um you know a really interesting Trend I've seen is you go to a website and you know when the web chat widget first launched as sort of like a paradigm I don't know like five six seven years ago um the Bots were not very powerful ful and so you would you know use the website then you use the widget for a little bit then you go back to the website the Bots are now getting so good that we've seen where like you click on the you know you click on the widget and it just goes full screen because there's no point in even using the website anymore and it starts to get you thinking like what what's the interface even going to look like like is web are websites obsolete now um as a paradigm because like I'm not even using the website I'm just using the chatbot and so you start to think well hold up like if the chatbot can start to give you know more detailed responses will when you land on a website page in future is it just going to be a chat bot and it will serve you the website elements as needed right and as you have a conversation with this website and so I think like that kind of interface uh Innovation is going to be so fast and Rapid that like we're going to let the community figure out how they want to do that and then just try to support it as best we can um so that's sort of the approach right now and then in future you can expect us to add more what we would call like direct Channel Integrations again so whether that's Instagram or Whatsapp but we just really want to focus on like enabling the community to innovate right now um and then over time I think we'll start to add more and more like native Channel Integrations it's funny you mention that as well because I was having a chat with my uh my template team on the the accelerator and had the same conversation I was just like racking my brain like where does this where does the interface go like how does this where does where does this all lead and what ises the interface of the future look like and how we interact with these assistants and Abby one of our one of our developers in there who was saying when like I I think the the interface becomes becomes chat in many ways and it does like you were saying make make the website in in many aspects obsolete because with the the ability to call functions throughout it as we saw with the DFD demo is that being able to converse with the assistant and be able to do things that you normally have to do by clicking around it's as I've said in Prior videos it's the end of this this point and click websites and it's trans sort of transforming over to a a conversational based interface but I think it's also important and a good an important reframe to make for people who are working in the space is that chat bots in the way that we know them have a have a basic and and sort of mundane and dull sound to them in many ways but yet we have it is literally the interface of the future at least before we get to the voice side of things I think it's been difficult pill for some people to swallow um it's something that I've been kind of having to push our pill along my journey is that people say oh but they wouldn't pay that for a chat bot it's just a chat bot it's just a this and based on conversations like this becomes clear that we need to reframe it a little bit as to how we look at chatbots it's not just a chatbot it's a it's a the new new way of interacting with technology in any ways so I'm glad to hear you you've got the same thoughts on that it is very exciting the thought of we've even considered to put it on our um on my on my accelerator website where it's okay you come on instead of having to read through the whole page people have questions that come to their mind as soon as they get to the page have to make them pick through the entire website to try to find it why not just ask the question and we can we can give you the answer and we can serve you up resources and links and everything that you need and I think the assistance API has been a big step towards that with the knowledge based component and the tools so I love the idea of reframing what a chat bot is um you know this is something when we were raising money for voice flow we had the same thing from investors you know who cares about chat Bots and you know the way that we frame this to folks is you know when you think about a chat bot you just think about like the chat bubble but it's so much more than that if you if you roll it back and you start to think about every single interaction that you have with a customer is a part of a conversation so when I'm on Netflix and I'm trying to choose a movie to watch and you know it's I've been scrolling for two minutes now that's me like you know in the most abstract way saying hey hey Netflix I can't find what I want can you jump in here and suggest something right like that was an implicit act in a conversation Netflix then jumps in over time you know uh I need to add another subscription I need to you know chat with them over text like all the different channels all the different interfaces all the different both implicit and explicit actions s to blend to uh chatbots essentially being like the CRM like the relationship operating system between uh the customer and your business and yeah and and like I think when people start to frame it that way that it is um the chat bubbles is the interface that people think about what people think that the interface is the product but the product is the agent the product is like that behind the scenes tracking all different interactions with all your customers and done really well it can be just an awesome experience uh like what if ever had it where you know you call an ivr and uh like a you know like a contact center and it already knows what you know it already knows what you're calling about you know because it looked up your number it saw you know who you are as a customer it saw that you purchased something last week like when you have experiences like that where again it feels like wow like this is not just some call center bot there's some intelligence behind the scenes like they have an agent there um that's when like the experiences get really magical people need to start thinking about disambiguating the interface that you see and the behind the scenes logic that is the operating system between how your business interacts with customers for sure I'm I'm very big on this this data side of things I think the the conversational a conversation as a channel via whatever the you want to use there's something special about being able to to pull information out of those conversations because you get to see into the customers world and life in a way that you don't really get to when they're just interacting with your website because they don't think that those conversations are being mon like it's not top of mind when you're asking something to a web chat widget about a product or about you're not thinking that this is going to be dug over and analyzed over time for for information and especially when it comes down to those more personal ones where it's not so much just customer support but it's actually like how can I help and it's getting very very personal with that and feedback as well there's there's so much to be minded out of this and I was looking up a good conversational data you might know something something like this but a like industrial scale or like business scale conversational data analytics that uh you can plug into something like these these assistants now as you're pumping out all these threads there needs to be a way I've just done a video on the channel um covering how you can export these threads and then analyze them um using air table to make it's nothing nothing too fancy but there's going to be massive need for okay I've got all these threads that have just been created over the past week where can I plug them in and get some insights out of it so do you know if there's any conversational data analytics platforms out there or or what do you guys use internally for that if you're allowed to say yeah I mean we um we've got some features and boic that aren't publicly available that do stuff like this and I'll maybe I'll show you after the recording um but I think a lot of people when they think about chat Bots again they're thinking on like the very first level which is like I see the chat Bubbles and I see it just for one particular use case like let's take lead gen but um I think what gets really interesting and I think this is for a lot of the um agency owners the very first step when you start working with a client and we're trying to do this more rigorously ourselves as well is let's analyze the data that you have let's analyze the transcript data like where is the opportunity for automation because you might think it's leg gen but it might be some internal uh task automation right and so for example you know we were chatting with a company the other day where um they had a bunch of uh I think they have like 600 cold call uh agents uh and what they need was uh lead enrichment and that was it like they just needed to be able to say hey I'm going to call this person run it through a chat bot that says hey I just pulled up older data here's like the best way to approach this cold call script M that by 600 and then like people 50 calls a day like 100 calls a day totally factoring the time of day factoring any potential holidays like and you start to realize that for that it's a very bespoke use case right leave lead jet is what I would consider um a very high volume high impact use case but where there's a lot of opportunity is um there's a thousand small little use cases at most businesses like you know we have use cases that only exist at Voice Low right things that you know only we uniquely need and how can we um work with companies that are automating not just like the core pillars everyone knows customer support everyone knows customer experience how can we start automating a lot of those bespoke um like mid mid volume low volume use cases that have high impact right so like we're starting to think a lot about this now is voo not just being a way to um you know automate customer support but it's like customer support's just the beginning once you're in you know once you've landed with a customer how do you then automate everything else right and really thing about like that AI automation piece which is why like I love sort of where um your s and others like are taking the space um because that's totally what it is it's AI automation but everyone needs everyone's focusing on the same use cases there's so many long use cases that provide a ton of value um and I think folks aren't necessarily thinking about it that way yeah I think that that comes down to me quite a bit I think I do try to at least in most of my content to make it as beginner friendly as possible and it does in order to attack those Niche use cases and the the business specific ones that gets a bit closer to a custom solution and in my experience you need a a fairly deep and and and Broad skill set as well in order to be able to piece together a solution for that so that includes like the automation platforms the chat Builders the the backend sort of more complex knowledge based functionality that you need so the I feel like the pillars are a good point for people to get into like a customer support leion and all these kind of areas where it's easy and and they understand it and there's places you can get templates like I've put lead generation templates and these kind of things out that are fairly easily attackable but that is the definitely the destination that we want to get a lot of the agency owners in the community to which is being able to come in there and and take a full sort of head to toe look at the business and identify those combination of of high impact and and low cost to deliver as well so it's it's ideally it's not ridiculously hard for you to build at the same time which is that sweet spot between valuable yet not too hard to deliver totally right like it's um and we think about this the exact same way um we have pillar use cases like you know our sales and marketing team very focused on customer support customer experience Automation in app co-pilots like you know sort of like the pillar use cases that people know about but what we're starting to do with uh our larger customers to say hey um if you're comfortable with it give us your call transcript data and let's see um you know you can do methods like utterance clustering let's look at the 10,000 utterances and let's cluster them into groups and you might realize hey there's actually this whole group of automation that's not even being touched right let's start looking at internal opportunities internal conversations um like conversational uh Analytics it's going to be a really big area for sure like um now that you can kind of automate everything like if you look at the history of chatbots over the past like you know 10 years in so of these know kind of came onto the scene in 2014 um chots were so expensive to build before llms that customer support was the only use case that even made sense from an Roi standpoint that was it it was like big companies had maybe their own custom teams like we work with a lot of big companies who you know they spend you know there you can go find some reports online $100 million building their chat bot is was a very big Bank in America that was how much they spent send my way next time they come around can deliver something there discount it's it's crazy so you see like a price tag like that I can tell you like from working with some of these teams there's hundreds of people like hundreds building these these Bots and a lot of them were built on the nlu Paradigm where that was what was required now with large language models you can have a two to three person agency stand something up that's like pretty comparable um and that's what's now leading to this explosion of like smbs uh you know midm markets being like hey like I don't need to have a $100 million budget to put together an assistant I can now do this with a very small team and that's what's been really exciting because it's now unlocking all these other use cases as well where again customer support was the only thing that had an Roi that made sense now it's like hold up if I can actually enable that you know 600 person call group you know and it's not going to cost me a million dollars to put that bot together that Roi is now actually worth it I'm going to go do it um and so I think we're just seeing like and I think it's like um the AI automation blown up we're just seeing like the proliferation of people realizing that the ROI statement that we've all been used to for the past 10 years has completely flipped everything is now for the most part like Roi positive as long as you have the right tools and skill set sure I think it'd be good to to also discuss what you've see to be as the the tech stack moving forward and how voice fits into that and yes we have the assistance API which is going to be running a lot of things we have some token costs that I want to talk to you about in a second about and the sort of usability of that in its current state but there's things like the assistance API there's voice flow and there's obviously the the major component of this which is tooling how do you see that sort of ecosystem tying together and how people can play into that and and build sort of solutions uh using your software yeah so how we think about voice flow is largely as an orchestration uh framework and so um how do you orchestrate multiple different models multiple different gpts um we're of the belief that uh if you want to build complex multi-turn um agents that are you know um reliably able to automate not just knowledge but tasks right because we sort of at voice how the internal framework we use is there's knowledge automation that's like your knowledge based Q&A like that's sort of that kind of stuff and then um there's the task automation how do we actually do refunds you know uh TI you know collect emails like that kind of stuff integration yeah essentially integration heavy stuff um we're very focused on okay cool you have you're gonna have gpts which are going to do a lot of the knowledge uh automation I think over time they'll get into some of the simple Integrations do we think open AI is going to be the only player in the space no so how do we be like the model agnostic orchestration framework that's really collaborative really powerful can scale all the way up from like you know a really uh Advanced agency all the way to the world's largest companies you know we have a bunch of features coming for scale soon um and that's how we're thinking about ourselves uh agnostic uh all theit orchestrating the different components because we don't think that for example let's say someone Builds an awesome GPT to do uh you know function a that's great have that like orchestrate with voice flow we've been showing off a lot of those types of capabilities recently and even as the gpts start to talk to each other we think that the ecosystem is going to be large enough where it'll never just be fully contained with open Ai and if it is that's totally cool at the end of the day we're just going to help customers build the best experiences but we believe in a world that's going to be a multimodel so there's gon to be a lot of different models out there you're going to want to keep uh some of your Integrations not just entirely with a no open AI you might want to have a third-party Integrations framework which is actually uh maybe you know hosted on a private cloud or whatever it might be um that's kind of how we're thinking about voice flow from from the long run um scalability extensibility um and then ultimately just trying to be agnostic to all the underlying models yeah I think we're on the same page there with just started work on our I'll probably have chat to you and Daniel about it a bit more and to get your thoughts on it but we are sort of rolling out our model agnostic platform that's going to help people to a lot of the agency owners to hook into platforms like voice flow and so it's sort of a little intermediary step between open Ai and and voice flow that's going to allow people to connect into it very easily and I'm super excited to get more information on about that but the uh I think we've seen with this open AI episode which turned our lives upside down I'm sure well it did for me I can only imagine for you with with the company you have on your hands um how much this past the past week must have felt but the the need for diversification away from why model provider has been really highlighted here for the past couple of days until Sam mman announced the the return I was thinking okay does this mean we've got if they're they're reallocating all their resources to be a research company again this means the assistance API is not going to get the development of needs it's it's still too slow it's still too expensive in many ways they haven't done memory management ET there's so much that needs to be done there but it was so promising it was really concerning like are we going to have to set and and TT thumbs here for 6 months until another company picks up the slack but thankfully we've now got them back on the scene however however that's that's going to play out but anord coming in with CLA 2.1 with tools I think is a very good sign for for what is to come and allowing us to be not so riant on just one provider so you said you have allowed and clawed support recently yeah yes yes we've got anthropic support uh all the open AI models actually I think I think we haven't launched turbo just yet but uh for gbd4 um yeah and anthropic um open a and then uh we've got plans to add a lot more models as well as custom models as well because we you know we work with some big Enterprises where they're just building their own stuff their own large language models uh and I think also with all the open source large language models uh coming out now it's I I think it's more likely than not that you have open AI remains kind of like the dominant close Source player and then you're gonna have just a slew especially on the big company side of like open source models um because it allows you control your data uh you know you don't have to worry about Cloud outage and things it's all in your Cloud it'll be really interesting to see how how much it fragments MH yeah I think tooling is going to be for for agency owners and people looking to make money in the space is to sort of ride the back of the stuff that's coming out what is your thoughts on uh tooling and the importance of knowing how to build their moving on because I feel like that out of the what I call like three ingredients out of these gbts and assistance you prompting knowledge and tooling any any idiot can do the the first two but being able to create custom tools that are powerful and specialized to a certain use case or adapted to a business I see is as what's going to be the biggest value add for someone who's looking to get into the space to make money how do you see the uh the tool creation environment continuing to develop and and the importance of it moving forward totally I actually want to answer this from a different angle and then I'll come back what I want to address is you know maybe a month ago there was uh I saw a whole bunch of videos it was like you know the the a AI AA model it's dead right all this kind of stuff you know I haven't made money um and I'll tie it back in why this is relevant in a sec but I thought it was such a like a ridiculous thing for this reason of think about how many like smbs have websites think about how many smbs um you know like it's it's you know and all of these smbs cannot use or it's really difficult to use AI uh today and so the way I view it is there's going to be a a massive Services layer in the middle I mean look at how deoe and all these like large consultancies are pushing into AI they're setting up practices or hiring You Know Chief AI officers and all this kind of stuff it's because if you do not have the skill set to stand stuff up that's the services layer in the middle right and so how we think about is Voice Low focus on building the best tools you're G to have all these smbs midm markets out there who do not have internal resources and it does not make sense for them to learn you know how to use voice loads we you know we use a website contractor because we don't have the time to use uh learn to use web flow to the degree we we need right it just doesn't make sense to hire a fulltime convenience and expertise you'll pay for it every time totally and so um I just think about how big the market is going to be for automating things with AI and all the smbs that will not have the skill set to do it and it's just like okay it's very clear there's going to be a large Services layer in the middle like that's just like it feels you look at the website market and I think AI is going to be wor impactful than websites you know because I think we're talking about not only uh the replacement of the interface which we've already kind of talked about but also the replacement of a lot of internal task so it's like the RPA market and the interface Market had like this you know this this yeah trying to find a good word but yes um and so anyways I I'm extremely bullish for that reason of like there's it just makes all the sense in the world so going back to how are good people uh people going to make money I think people are going to make money setting up um a gencies and I think when people think about agencies uh you know I mean you've got a ton of great resources on this but like choosing a use case for sure like building up like building up an expertise and overtime branching out to tental use cases that's awesome because I think um customers and smbs don't search for AI automation agency they search for lead gen automation you know I mean they search for the use case you just happen to be the model and so I think some people have got that you know kind of backwards thinking that people are searching for the agency name yeah no one outside the space knows what AI automation the agency is no no business owner is searching that up totally like we don't you know we're not trying to optimize SEO for like uh you know like collaborative chatbot Builder we're Auto you know we're doing like customer support automation because people are searching for the use case that the tool solves for so that you know that's sort of the first thing so agencies with use cases then I think you're G to have that's all for the video here on the main channel the rest of it is going to be in its uncut full link format here on the morning side AI channel so if you want to watch it you can click up there if you have enjoyed please hit down below and leave a like and let me know if you want to see more content like this and if you haven't already you can subscribe to the channel for more AI business content