AI Innovators · AI Dev Lab
AI Dev Lab  ·  Impact Series

AI
Innovators

The builders doing something genuinely new with AI.
Not hype. Not demos. Not another benchmark. These are the people who saw a real problem, ignored the consensus, and built something that actually works.

Meet the innovators
Featured this series
Dylan Fox
AssemblyAI
Audio AI
David Collen
SapientX
Conversational
Arjun Rai
helloWoofy
AI Marketing
Tobias Martens
Whoelse AI
Voice AI
About This Series

What makes an AI innovator
worth paying attention to?

Most AI coverage follows the same pattern. A new model drops, the coverage spikes, and then everyone moves on. The builders doing serious work rarely get much attention — because what they're doing is hard to explain in a headline.

AI Innovators is our attempt to fix that. Over the years, I've had conversations with founders who were doing something genuinely new — not in a press-release sense, but in the sense that they were solving real problems others had decided were too hard.

Every person featured here was selected for one reason: the work speaks for itself.

Selection Criteria
01
They solved a specific problem first.
Specific, underserved, clearly defined — before platform or scale.
02
They questioned the dominant approach.
Even when the consensus was backed by billions of dollars and thousands of engineers.
03
They kept the human at the center.
The technology was always in service of someone's actual experience — not the other way around.
“
"These conversations shaped how I think about what makes AI actually useful. That thinking is what I'm sharing here."

Jason Wells  ·  Co-Founder, AI Dev Lab

What the best builders have in common

Three patterns worth paying attention to

After years in this space — across voice AI, audio intelligence, marketing automation, and conversational systems — the same signals keep showing up in the work that actually succeeds.

01
Specificity First
Solve a specific problem before building a platform.

Dylan Fox saw developers couldn't get reliable speech recognition. Arjun Rai noticed small businesses were losing marketing battles they shouldn't. David Collen identified that most conversational AI was built the wrong way. Specificity came first. Ambition came later.

02
Challenge Consensus
Question the dominant approach — even when it's expensive to be wrong.

Each of these founders had a moment where they looked at how incumbents were doing things and decided there was a better path. That willingness — even when backed by billions — is what makes genuinely differentiated work possible.

03
Human at the Center
Keep the person on the other end at the center of every decision.

Whether it was a small business owner, a frustrated developer, or a senior citizen who wanted a voice assistant that felt like a real conversation — the best AI builders stay close to the actual user. Technology is always in service of something human.

The Series

Four builders. Four domains.
One standard.

Each profile extracts the principles and decisions that stay relevant — regardless of when you read them.

Dylan Fox, Founder of AssemblyAI
PROFILE 01Audio Intelligence  ·  AssemblyAI
Dylan Fox
Founder & CEO, AssemblyAI

How a developer-first approach to audio intelligence built a category that Big Tech missed. Dylan left Cisco when he saw a clear gap: developers needed reliable, well-documented speech recognition, and nobody was building it for them. What he learned about infrastructure, support, and trust is a lesson for every organization betting on AI tools they don't control.

"What if you could build a Twilio-style company using the latest deep learning research — one that's actually built for the developer, not the enterprise sales team?"

Build vs. Rely  ·  Developer Experience  ·  Infrastructure Risk Read the profile
David Collen, Founder of SapientX
PROFILE 02Conversational AI  ·  SapientX
David Collen
Founder, SapientX

Why most conversational AI underperforms — and what a radically different technical approach reveals. David put the first 3D model on the internet in the 90s, built soldier tracking systems for the Army, and developed early voice assistants for automotive navigation. His take on symbolic reasoning vs. machine learning cuts through the orthodoxy.

"Machine learning is not the be-all and end-all. If you crack open an AI textbook, there are many different branches. Some approaches are simply better at certain things."

AI Architecture  ·  Conversational Design  ·  Practical AI Read the profile
Arjun Rai, Founder of helloWoofy
PROFILE 03AI Marketing  ·  helloWoofy
Arjun Rai
Founder & CEO, helloWoofy

Using AI to level the marketing playing field for small businesses. Arjun's central argument: small businesses aren't losing marketing battles because they lack hustle — they're losing because they don't have access to the same data, tools, and intelligence that large companies take for granted. He analyzed hundreds of millions of phrase combinations so a business owner can just start typing.

"We believed that AI algorithms could level the playing field — that digital marketing underdogs could compete equally. That belief became the entire company."

AI Democratization  ·  Domain Knowledge  ·  Small Business Read the profile
Tobias Martens, Founder of Whoelse AI
PROFILE 04Voice AI  ·  Whoelse AI
Tobias Martens
Founder, Whoelse AI

The fragmentation problem in voice AI — and why interoperability matters more than any individual feature. There were over a thousand voice AI technologies on the market when Tobias and I spoke. That gap isn't a marketing problem. It's a standards problem. Tobias came from European Commission policy and ISO standardization work, giving him a completely different lens on what slows adoption.

"We're not seeking the greatest standard currently available. We want the most basic standard feasible — built for linguistic structure, nothing more."

Interoperability  ·  AI Standards  ·  Infrastructure Read the profile
Why This Series Matters

The same principles that drove these innovations drive how we work.

If your organization is evaluating AI tools, building AI products, or trying to figure out where AI fits — these profiles are useful beyond the individual companies featured.

They show what it looks like when someone gets it right. More importantly, they show why. The principles that drove success for an AI startup are the same ones that drive successful AI adoption in a transit agency or service organization today.

Find a specific problem. Question the default approach. Keep the human experience at the center. That's what we bring to every engagement at AI Dev Lab.

Ask whether the vendor you're evaluating treats your use case as core or peripheral. The answer tells you what happens when something breaks.

Domain knowledge baked into AI beats generic AI pointed at a domain. The tools that win in specialized contexts were built by people who understood the context first.

AI adoption stalls not because the technology doesn't work, but because it asks too much before the user sees value. Reduce friction before adding features.

Before committing to AI infrastructure, understand how tightly you'll couple to a single vendor. Fragmentation is a real cost — design around it early.

Work with a team that
knows the landscape.

The principles these innovators followed are the same ones we bring to every AI engagement. If you're building AI products or figuring out where AI fits in your organization — we should talk.

Let's Talk No commitment  ·  30 minutes  ·  Senior leadership
Jason Wells
Jason Wells
Co-Founder & Chief Strategy Officer, AI Dev Lab
MBA, Wharton MS Applied Mathematics Former SVP, Sony Pictures Kearney Alum 4× Ironman
Building AI products in transit and enterprise since before it was a pitch deck category.