Twilio Signal 2026 | Evan Mendenhall, Professional AI Agents, LLC

Clips
More from Twilio Signal 2026

Evan Mendenhall

Founder

Professional AI Agents, LLC

play_circle_outline Introduction: Twilio SIGNAL 2026 interview with Evan Mendenhall, founder Professional AI Agents.

play_circle_outline Twilio edge data centers simplify model hosting and GPU procurement.

play_circle_outline Evan Built AI Agents in Early 2024, Placed Top-100 in NVIDIA Contest, Now Runs Agent Services Company

Info
Transcript

Evan Mendenhall, Professional AI Agents, LLC

Evan Mendenhall

Founder Professional AI Agents, LLC

In this interview from Twilio Signal 2026, Evan Mendenhall, founder of Professional AI Agents, joins theCUBE Research's Paul Nashawaty to discuss Twilio's evolution from a fragmented toolset to a unified AI platform and what that shift means for the next generation of intelligent agent development. Mendenhall, who has been building AI agents since early 2024 and placed in the top 100 of NVIDIA's global AI agent developer contest, offers a builder's perspective on the six product launches unveiled at the SIGNAL keynote. He highlights how Twilio's Conversationa... Read more

explore Keep Exploring

Please introduce yourself. add

How realistic does the "marketecture" slide showing how all the pieces (conversational relay, hosted models at the edge, integrated voice/ASR components, etc.) come together seem to you? add

Can you describe your experience placing in NVIDIA's AI Agent Developer Contest and how that influenced your decision to found a company? add

bolt Powered by CUBE AI

Evan Mendenhall, Professional AI Agents, LLC

search

Paul Nashawaty

>> Hello and welcome back to Twilio SIGNAL 2026. My name is Paul Nashawaty. I'm coming to you live from San Francisco. I'm here with Evan Mendenhall, founder of Professional AI Agents LLC. How you doing?

Evan Mendenhall

>> Hello. I'm doing great. It's great to be here.

Paul Nashawaty

>> Great to have you on. Well, let's start with, why don't you introduce yourself?

Evan Mendenhall

>> Yeah. Well, I'm Evan Mendenhall, as you said. I did found Professional AI Agents, which ended up being quite the right name to choose for sure a couple of years ago. So I've been in the space building agents basically in the middle and early 2024.

Paul Nashawaty

>> Okay.

Evan Mendenhall

>> And now they're relevant and needed everywhere.

Paul Nashawaty

>> Yeah, absolutely. So you were at the forefront of all of this while it was just coming together.

Evan Mendenhall

>> It seems like it.

Paul Nashawaty

>> Yeah, absolutely. Well, okay. So we're here at Twilio SIGNAL. You're a Twilio champion. You've been building a lot with Twilio projects. What do you think of all these announcements? I mean, the change from year over year has been dramatic.

Evan Mendenhall

>> Yeah. It's incredible to see the roadmap that you guys have released. I think it was six product launches today in the keynote. And a lot of them are actually going to make my workflow a lot easier. So I'm pretty excited to migrate some of my legacy systems over.

Paul Nashawaty

>> Yeah, that's awesome. I mean, I know that even a year ago, all the products were pretty much in silos. You could do it, you can build it, you can put it together.

Evan Mendenhall

>> Sure.

Paul Nashawaty

>> But it was like a bag of bits.

Evan Mendenhall

>> Same grid. You have to go find it. I like that it's pulling the console now.

Paul Nashawaty

>> Absolutely. There's one platform view, one experience, one platform, one view. And then all of the announcements, whether it's orchestration into memory, into collaboration, communication, you said it's going to make your life easier. How so?

Evan Mendenhall

>> Yeah. Well, I'd like to basically use the conversational relay to house a lot of the pieces that I had to configure myself in the past. So typically I would just build each piece through a Python API integration, pull in Deepgram for this particular aspect of it, use a text to AI voice in another section. But especially it seems like with conversational relay, you guys have a data center that's hosting a lot of the models that I would rely on, especially the smaller, faster models. And so one of my big challenges, especially for voice agents, has been mapping, tracking, and then reducing the millisecond latency across 10 or 15 different pieces of the tech stack. And it's been a major challenge, but inside conversational relay, as long as it's on that edge as it showed on the map today in the keynote, many of those pieces will all be in the same data center. So that's just going to, I mean, that saves tons of headaches and probably shaves off 300 more milliseconds.

Paul Nashawaty

>> So okay. So from your perspective, co put up the, what I would call like a marketecture slide. About how all these pieces come together. How real is that to you?

Evan Mendenhall

>> It seems pretty good to me. In fact, yeah, I've been working with several other founders because this is just an exciting time in the startup space. And so there's a lot of really good companies that are ambitiously attempting to do this. And they're a good bit behind what I saw with the map today. And so it usually comes down to seeing if they can get all of these deals for GPUs, locate them, and then how do you get the models that are usable enough for agents in order to actually house them? So they've run into some pretty big problems that it seems like you guys are basically handled as of today.

Paul Nashawaty

>> Yeah. Yeah. Well, I mean, that's coming from you. It's actually very interesting. You were placed top 100 globally in NVIDIA's AI agent developer contest. Tell us a little bit about that. That sounds really cool.

Evan Mendenhall

>> Yeah. I would say it sounds better than it is. I don't know if there are millions of people building agents three years ago or something. So it was probably a smaller contest pool. But yeah, it wasn't first or second place. It was top 100. I got a science certificate from Jensen, which was pretty exciting. But basically what I wanted to do was validate my skillset. I was looking to find out how much of this do I really understand? And I really want that feedback from the experts in NVIDIA. So placing it all shocked me. And when I saw that, that was the realization that, okay, I guess I understand the pieces enough that I want to found a company and I'd like to provide this service for other businesses because you mentioned I was early. I mean, people have been studying AI for decades and decades, but it felt like now is the time that every business needed what I now knew how to build.

Paul Nashawaty

>> Yeah. Well, I wouldn't undersell it. I mean, you definitely did some early work there. And even though you may not have had the same volume of people that are there today, it's still it's a quite of achievement.

Evan Mendenhall

>> I appreciate that, Paul. Thanks.

Paul Nashawaty

>> Yeah, absolutely. Well, I do think that it's interesting because your journey is a lot of what organizations are going on in their journey today. Whether they're early in their maturity and their journey is just trying to just set take off. There's outliers on both sides. There's people that are very organizations that are very mature and they know how to use this very well. But then there's people that are like, "Hey, this is really scary. I don't even know what to do with this."

Evan Mendenhall

>> Sure. We see a lot of that.

Paul Nashawaty

>> Yeah. What would you advise those people if you-

Evan Mendenhall

>> I don't think I really have any advice for them. And that's basically because America is amazing where truly the founder and whoever it is that has this idea, they can build the company that represents what's important to them. And many companies aren't the fastest to adopt tech necessarily, but they really like their traditional way. And that is going to be a problem and they don't have the competitive edge, but I wouldn't say I would have certain advice to people who are resistant to it. On the other side of things, those that want to adopt, which is the large majority of us, all of the standard businesses, we do want to adopt it. It's just, we need to realize that it's almost cliche of how fast it's moving, but we actually need an agentic layer of research and recon in order to find out what's happening fast enough so that we ourselves can stay up on the tools being used now. And that is one of the things I found to be most interesting.

Paul Nashawaty

>> No, it's very fair. I mean, I ran my 2025 research and I found that 50% of code as of August was using AI generated code. I reran that research at the end of 2025, so November, December, and jumped up to 70 to 90% of-

Evan Mendenhall

>> Exactly. Yeah. It's so high right now. It's incredible.

Paul Nashawaty

>> It is. And I think there's two things. One, 2025 was a year of innovation and experimentation. '26 is the year of implementation. And it's like we're trying to get these things that we tried and build it and put into production, you got to get real. And it's showing that people are more in tune to how AI is working and trusting it. That was a big message that we were seeing at the show, trust.

Evan Mendenhall

>> Big fear in the beginning, for sure. There's a faster adoption curve. And I think for better or for worse, I think OpenClaw was one of the reasons that it got pushed so far ahead so fast and it was just so quickly, I mean, it was I think 300,000 GitHub stars or something faster than anything in history. Linux got 300,000 stars, which the whole internet's based off of over 30 years. And so now I think in a very poor way, I don't really love the trend there. I think people are just handing their lives over to AI and say, "Here's my credit card, here's my personal information, here's root access over my hardware, please live my life for me." And so the best AI engineers I hear don't use OpenClaw, but it was so fantastic in bringing the masses to say, "Whoa, AI can actually do a lot more than I thought."

Paul Nashawaty

>> Yeah, I agree with that statement, but let me ask you this, your thought on this. Is it a pivot of letting AI take over? Is it laziness or is it the fear of being competitively boxed out or what's the challenge here? Why are people so willing to give up their identity almost to AI?

Evan Mendenhall

>> Yeah, it's pretty interesting. There's a large portion of laziness in the sense where it's just like, "Okay, please start a business for me and come up with the idea to build that business and then manage that business and then find me clients and okay, cool, back to Fortnite." But the other sense is, yeah, essentially there's a hesitancy because those that have been in the coding space much longer than me, they essentially, they have this expertise that they're proud of and they should be proud of it. It was very, very difficult to learn the Python that I learned. It is a different language. You do learn to speak it. And so there's a hesitancy to say, "Wow, my own skillset has been replaced." I knew AI was going to replace the rest of the jobs and now my coding job has been replaced. Andrew Ng actually just said at AI Dev26 that he has his code basically checked by an AI and that now the main goal is to get yourself out of the way, let more coding happen by coding agents and let the code get checked by agents. And if you're in there, you're the bottleneck. And when I heard Andrew Ng, who is the top 100 influential people in AI, led the Google brain team, I took his Stanford machine learning classes online, that's how we met. But once he was willing to say, "Okay, I'm going to put down my ego and let code build code." I had to say, "Well, why am I a holdout?"

Paul Nashawaty

>> Yeah, I mean, it's fair. I mean, look, in my career, I've built assembler code and then we moved up to C code and then we moved up to all the different evolution.

Evan Mendenhall

>> That's exciting. Okay.

Paul Nashawaty

>> So you know it from the machine elements to all the way up to the logic business logic, but that's an evolution. And you know what? I would never go back to writing machine code at this point because-

Evan Mendenhall

>> Yeah, it doesn't make sense. We've abstracted that away.

Paul Nashawaty

>> There was a time and place for it, but using natural language and compiling based on what an AI can do, by all means that's what we do. It's an evolution. We want to get to that point. I mean, I think that makes a lot of sense. So based on what you're seeing here at Twilio, this evolution of product silos that they had just a few years ago to now a one platform view. That makes things more streamlined, the rapid time to value, or I was on earlier with a couple of the Twilio folks and they were talking about the rapid time to magic and then they were talking about that.

Evan Mendenhall

>> Yeah, exactly.

Paul Nashawaty

>> I think it's interesting. What are your thoughts on that?

Evan Mendenhall

>> Well, yeah, removing the silos basically brings down the learning barriers. And so obviously learning any code in the past just took many years of education for each piece of it and now truly we can all code at the speed of thought. So if we have a platform now that unifies all of these different pieces, the infrastructure that all of this great communication platforms and services are built upon, then truly anybody can be a coder. And I think that's just an exciting time. And I'd love for that message to get through to people who don't think they can do it, because if they can think, they can code now with the assistance of a coding agent.

Paul Nashawaty

>> No, I agree. I think that we've seen this kind of bifurcation in the developer world. Developers are builders. Lines of business are enabled to be builders. So you can have sales, marketing, finance, they can build their applications using AI and have a professional developer focused on the true innovation. I mean, the age-old argument. I don't want to be maintaining my code. I want to innovate. And now you're into like, now they can actually do that because they don't have to spend their time getting servicing requests from sales, marketing, and finance. They can build and innovate as they need to. So I think it's a good shift. I think watching what we saw at the keynote today, watching the one platform view, the unification of all the different products coming into the platform makes it more interesting to the lines of business, but it also makes it accessible to your point of we can have different people using the platform that weren't able to use it in the past.

Evan Mendenhall

>> And as you were mentioning, I mean, it just brings that friction down where you're not proofreading, reading, checking, debugging as much as you were in the past, especially at the speed of AI checking code. One of my friends, he does a PhD in math. I've known him since we were both 10 years old. He's wanted me to learn Python my whole life. And so he's finally excited that in the last half decade I started studying it. So he's been able to essentially take on a data science role in a company here in San Francisco where he was basically facilitating and helping junior data scientists. And basically what he said to me was that he's been able to do five times the data science for each of the five and one fifth of the time. I'm like, "Look, I know you're a math guy. I'm not going to try to trace that one, but okay, you're more productive."

Paul Nashawaty

>> It tracks though. I totally get where he's going with it.

Evan Mendenhall

>> He's excited because in the past he would have to sit there debugging or read the code, but now he can really just orchestrate it, think it through and build the framework with AI.

Paul Nashawaty

>> Yeah. I think that's a really good way to think about it and on how it's coming together, on how it's being more productive. So, all right. So Evan, I think that you've done quite a bit with Twilio. You've built a successful business around it. You have your head focus in the right direction of where the market's going. I think that's really awesome. Where do you see it? If we were sitting here in 2027 at that Twilio SIGNAL 2027, where do you see us going in the next year?

Evan Mendenhall

>> That's a great question and I really can't tell. And really my stance of it is I think we're at a pace of such exponential and logarithmic change that really no one can forecast what's about to happen. And so that's a scary and interesting time. And so I have no idea. Even that short term, I really don't. But what I think is we can prepare to it, prepare by learning every single day on what's changing.

Paul Nashawaty

>> Yeah, I agree. I think that it's anyone's guess at this point. You can see what happens. Well, Evan, thank you for being my guest today. This has really been insightful and congratulations on all your success.

Evan Mendenhall

>> Thank you.

Paul Nashawaty

>> This has really been, really been great.

Evan Mendenhall

>> Yeah, of course. Happy to be here. And I really, all of my systems have been so much better with Twilio. My best product is a voice agent over self-storage facility management basically. And I brought in the Twilio lines, which just reduced the dropped calls, it reduced the static. And so I fully plan to migrate more of my stack directly on the Twilio platform.

Paul Nashawaty

>> Excellent. Well, I look forward to seeing that as it comes together and thank you. And thank you for watching us. Paul Nashawaty coming to you live from Twilio SIGNAL 2026, coming to you from theCUBE, your leading source of tech news.