AI is moving at a pace that leaves Moore’s Law — the idea that computing power doubles about every two years — in the dust. And many believe that the potential for some of humanity's biggest milestones are just around the corner with AI capabilities that have emerged in only the last year.
Take the robot made by Amazon that can now physically feel items that it picks up due to AI. Or the OpenAI large language model that passed the Turing Test 73 percent of the time.
Or the dawn of quantized intelligence, when quantum computing and AI are combined.
But AI advances like these don't exist in a vacuum. Building the newest AI models, AI agents, or even using AI tools to transform workflows, all require innovative engineering teams that are willing to push the boundaries of what we assume is possible.
Built In spoke with several New York City tech leaders to hear about their strategic approaches to AI research and development, data management and how they create thriving team cultures.
Rokt is a global leader in e-commerce, unlocking real-time relevance in the moment that matters most.
What is the unique story that you feel your company has with AI? If you were writing about it, what would the title of your blog be?
If I had a blog, I’d probably give this article a title with a corny pun like: “AI Everywhere All at Once.” The story goes, one day, everyone at Rokt woke up and had access to some new AI tool. Shortly after, we all woke up to access instructions for another one, and shortly after that there was even more. The cycle continued and suddenly everyone was just a few keystrokes away from the best tools the industry has to offer right now. Rather than deciding which tools should be licensed, Rokt decided to give everyone access to everything and let each individual decide where the best allocation of value is.
The collaboration across the organization that made this shift feel so sudden was truly a sight to see. I’ve never seen a collective effort to democratize access to tooling and information like this in my career and I’m proud to work with people so passionate about leveling the playing field for experimentation and upskilling across the entire organization. Legal, privacy, access management and a multitude of other considerations were handled with speed and care, each with the objective of enabling the individual to do more.
What are you most excited about in the field of AI right now?
I’m fascinated by how much individuals can now learn as the time and resources needed to fail decrease.
The quicker someone can break something, uncover a bias or test a rough solution, the quicker they can solve for the realities they face. Staying critical and curious is essential as we navigate new paths to information and rapid prototyping — we have everything to learn when nothing goes as planned.
What adds to that fascination is the scale of accelerated iteration. With just an internet connection, people can access powerful tools that fuel creativity and redefine how we build technology. As learning curves flatten, more individuals can take innovation into their own hands. I’m excited to see what creative minds from anywhere in the world will build.
From an industry perspective, as monetary barriers fall, I’m interested in how value transfer evolves. With more individuals able to craft bespoke solutions, even industries without strong technical foundations have an opportunity to modernize through experimentation, learning and speed.
How do you learn from one another and collaborate?
Continuous learning on my team is a byproduct of each person’s commitment to creating a safe, level environment for communication.
Most days, you’ll find teammates gathered in small groups solving problems we’ve never seen before — huddled around a desk investigating anomalies or debating solutions to issues in our processes or documentation. We support each other, ask intentional questions and spend real time trying to understand one another. That foundation of trust is what allows us to navigate the unknown — together. We share what we learn with sincerity and with a genuine drive to lift each other up, constantly pushing into new frontiers as a team. It’s a formula for resilience in a world where how we work is changing fast. As the mediums for collaboration and communication evolve, we rely on each other — not hierarchy — to move forward. The new wave of AI tools is just another unknown we’re solving for side by side, with emphasis on together.
Zocdoc is a tech company that enables millions of patients to find in-network neighborhood doctors, instantly book appointments online and manage their healthcare experience.
What is the unique story that you feel your company has with AI? If you were writing about it, what would the title of your blog be?
I’d title our blog “Human Outcomes, Machine Support.”
At the forefront of the AI revolution is how it’ll impact us as humans — how will AI affect jobs, redefine productivity or alter our sense of self? Zocdoc, a company built around patient experience, sits at a unique intersection of AI and human care.
Our journey with AI is just beginning, but already it’s reshaped our business. Our first major endeavor with AI expanded our healthcare scheduling expertise to America’s highest-volume scheduling channel: the phone. We launched Zo, an AI phone assistant that instantly schedules appointments 24/7 using natural, conversational language. This initiative, championed by Chief AI Officer and co-founder Nick Ganju, aims to remove friction from the phone, help patients seamlessly schedule care and soon order prescription refills, get referrals and handle other needs.
Looking beyond one feature, we’re building AI fluency across the company. From recurring AI knowledge shares to active Slack channels, we’re cultivating a culture that’s curious and responsible. Yet, we’ve never lost sight of the patient. The drive to give power to the patient is what makes Zocdoc a unique and inspiring place to shape the future of healthcare.
What are you most excited about in the field of AI right now?
The open source space excites me the most right now. Having this technology in both the hands of big tech companies as well as tinkerers expands the breadth and depth of what can be accomplished. Yesterday, my fellow engineers and I were the builders; today every person reading this is a builder. When the number of people able to build and design increases exponentially, the rate of progress becomes breathtaking.
This AI technology doesn’t just make development accessible, it changes the foundation of how we interact with machines. For decades, we had to think like computers to build with them. Now, machines are learning to think more like us. Language has become the interface. That shift opens the door to collaboration between technical and non-technical minds in a way we’ve never seen — doctors designing patient flows, researchers automating insights, creatives prototyping ideas in real time. The barrier to creation isn’t just lower; it’s dissolving. While companies like OpenAI and Anthropic are leading the way, it enables non-technical people across practices to stand on those shoulders and imagine what’s next.
How do you learn from one another and collaborate?
The surprising thing is: learning AI itself isn’t the hardest part. With so many resources available — blogs, videos and courses — and with the tech rooted in natural language, the basics are more accessible than ever. The real challenge was diving into unfamiliar domains needed to make AI useful in the real world.
For instance, building an AI phone assistant requires a great deal of engineering, architecting and designing. For a team of highly skilled engineers, it’s not an insurmountable task with our existing skill set. However, learning about voice UX was an unforeseen complication that I didn’t realize would be a hurdle. Design a framework that will be robust and low latency? Not a problem. Figuring out how to phrase a certain question so that the fewest number of people are confused by it? My team and I had to start from square one.
That’s why continuous learning is part of our daily workflow. We spend time not just building, but sharing learnings, mistakes and insights. This cross-disciplinary learning is becoming the new normal in AI work. Success isn’t just about understanding models, but understanding the real-world systems and human contexts it interacts with.
Regal is an AI platform that empowers companies to transform customer communications with AI agents.
What is the unique story that you feel your company has with AI? If you were writing about it, what would the title of your blog be?
Blog: “From Human Agents to AI: How Regal Scaled to Meet the Future of Customer Conversations.” Regal didn’t start as an AI company. We began as a modern contact center solution for high-consideration industries like insurance and healthcare, born from a simple insight: ads drive traffic but conversations close deals. We built for that conversation, combining campaign orchestration, multi-channel messaging and real-time data to help agents reach the right lead at the right time.
As we scaled, we saw the same challenges: limited agent capacity, high overhead and long wait times. We knew AI could solve this, but the tech wasn’t ready. Then it was. Voice quality, LLMs and latency reached a tipping point, and what we had seen coming became clear to the market: voice was the future and AI agents were crucial.
Today, Regal’s AI agents drive real outcomes: increasing reach by 40 percent, doubling conversions and resolving 100 percent of inbound calls 24/7. Built on our orchestration engine, they adapt in real time, follow up across channels and hand off to humans when it matters. What sets our story apart isn’t that we adopted AI, it’s that we were already built for it.
What are you most excited about in the field of AI right now?
As a product person, I’m most excited to define best practices for building and monitoring AI agents. We’re not trying to catch up to an already defined ideal — we’re creating it. The ambiguity is a challenge; customers don’t know what they need, and no one has the blueprint yet, but it also opens up a world of innovation.
On the build side, we’re focused on empowering non-technical users to effectively prompt and configure their AI agents. We’re teaching them how to fine-tune voice settings such as tone, pace and background noise, and how to structure and test prompts to launch quickly and effectively.
On the monitoring side, we’re defining which metrics matter (e.g., containment rate, success rate and CSAT). Since we own the full data stack across customer profiles, actions and conversations, we’re in a unique position to surface insights our customers wouldn’t know to look for. That’s where I see the biggest unlock: using this foundation to answer questions like which agent voice works best for which audience or which phrases signal low versus high intent. We’re building toward real-time insights to help teams test, learn and optimize with confidence.
How do you learn from one another and collaborate?
Continuous learning starts with staying close to our customers. Even our co-founder is in the field, prompting AI agents, monitoring performance and making real-time tweaks. That level of hands-on engagement helps us understand what’s working, what’s not and where the most significant opportunities are.
We prioritize working with customers who are data-driven and open to experimentation so we can test, learn and iterate together. Our platform’s built-in experimentation toolkit allows us to move fast while minimizing risk for sensitive brands or use cases. We can ship a change to just 1 percent of calls, validate its impact and scale it gradually.
On the development side, we stay nimble, running proofs of concept to assess risk and lift before committing to a full feature set. We discuss complex decisions in impromptu chats with the necessary stakeholders, rather than scheduling recurring meetings. Knowledge-sharing happens constantly via Slack, training and a dedicated budget for taking AI courses. We encourage non-engineers to build their own AI agents, both to validate that the experience is intuitive for non-technical users and to help us catch bugs or usability gaps early.
Fora Travel is a travel tech company that offers a suite of tools for booking, client management, marketing and more.
What is the unique story that you feel your company has with AI? If you were writing about it, what would the title of your blog be?
One of the most common things I’ve been asked about AI since joining Fora Travel by people outside the company is if we are going to replace travel agents with AI.
This is top of mind for everyone these days, and everyone is scared of how they will be impacted by AI. While AI isn’t broadly replacing human jobs yet, it is only a matter of time before many repetitive, mundane aspects of people’s work will be automated. Ideally they will be freed up to work on things that require a human touch. Fora is uniquely focused on empowering the next generation of travel entrepreneurs — not to replace them like online travel agencies are trying to do. Our edge is building an unstoppable combination of travel advisors supercharged by the kinds of AI tools only we can build.
Travel advising is a job where the human touch will always have the upper hand. Imagine you’re on the other side of the world and you are having a big problem with your travel plans. Who would you rather talk to? an AI chatbot? a random customer service rep at some huge corporation? or your dedicated travel advisor — someone who knows you, your itinerary, your preferences and has been with you every step of the journey?
What are you most excited about in the field of AI right now?
I’m most excited about AI coding assistance, particularly in the context of building prototypes from scratch, because it really works today and it’s like magic. The current buzzwords are “zero-to-one” and “vibe coding.”
We recently had a hackathon at Fora and it was one of the most productive hackathons I’ve ever seen. Every project felt like something that we could launch as a real product. We even had some teams with minimal coding experience build impressive, functional prototypes with the help of AI — and the prototypes themselves used AI services.
Of course, the elephant in the room is the automation question: Will software engineering disappear? I don’t think so. It will evolve, just like software engineering has always been evolving. In the old days, coding involved thinking about low-level details of how the computer works. But over time, programming has gotten more abstract and closer to natural language like English. Tools change, and people who work in cutting-edge technologies must continually adapt. I’m optimistic overall, but the challenge will be in finding the right balance between leveraging AI and not outsourcing human creativity and critical thinking.
How do you learn from one another and collaborate?
I’ve been in the field for about 25 years, so I guess I’m one of those people with years of experience to draw from. But no one — regardless of experience — has seen the current pace of change in AI. It’s truly historic.
Continuous learning means being curious and humble. AI is mainstream now, and many updates are in the mainstream media. This means that you may hear about a new model or technical paper from anyone, so you need to be open to learning from anyone and everyone.
The book is not written on how to use generative AI to get the best results — it’s being written now, in real time, and we’re all learning by doing it. So it’s essential to make time to talk to others and learn how they’re using these tools.
At Fora, this means I’m regularly connecting with travel advisors and team members across the company — not just the AI and engineering teams, but everyone from marketing to sales to finance to legal and everything in between.
Outside of Fora, it means connecting with my colleagues, networking and attending conferences and webinars. The only way to be successful with the deluge of new research, models and techniques is to leverage the power of your social network.
GameChanger is a sports tech company that helps families elevate the next generation through youth sports.
What is the unique story that you feel your company has with AI? If you were writing about it, what would the title of your blog be?
Our early work in downtime detection focused on recognizing the rhythm of youth sports, pinpointing when play starts, pauses and shifts. These weren’t polished, high-budget broadcasts. They were handheld cameras, unpredictable framing and all the beautiful chaos that comes with youth games.
As one of the earliest hires on the computer vision team, I helped build the models and infrastructure that made this possible. That foundation challenged us to move beyond traditional rules and build systems that could adapt to real-world conditions.
Today, we’re taking that even further. Our action recognition work now focuses on helping AI interpret how the game unfolds, not just when. It’s about watching the game like a coach, not a camera.
What are you most excited about in the field of AI right now?
A major milestone came when our downtime detection models could reliably trim full-length games down to only the moments that matter, at scale and in the wild. That meant our AI wasn’t just working; it was interpreting.
I was directly responsible for building the model architecture and training pipeline that made this possible. After months of iteration and experimentation, seeing our system accurately cut out downtime across thousands of videos felt like watching the model understand the game for the first time.
Now, what excites me most is pushing our action recognition even further, moving toward systems that grasp the tempo, intensity and tactical nuance of each play across different sports. What we’ve built so far is just the beginning. There’s a whole wave of AI-driven features we’re preparing to roll out.
How do you learn from one another and collaborate?
One of the greatest challenges we faced was the sheer variability of youth sports video — shaky footage, inconsistent angles and unpredictable pacing. Our downtime detection and action recognition models had to adapt to all of it.
What started as scrappy prototyping evolved into robust, production-scale systems, thanks to close collaboration between CV, engineering and design. We still operate with a startup mindset: fast iteration, open feedback loops and constant refinement.
We treat our models like living products, constantly learning and evolving based on real-world data. Continuous learning isn’t just about staying current with research — it’s about staying close to our users, our data and each other.
Fluent is a commerce media solutions provider connecting top-tier brands with engaged consumers. Leveraging diverse ad inventory, robust first-party data and proprietary machine learning, Fluent unlocks additional revenue streams for partners and empowers advertisers to acquire customers at scale.
What is the unique story that you feel your company has with AI? If you were writing about it, what would the title of your blog be?
The title would be "Aim High," and it would tell the story of a small, focused tech team looking to make a big impact in the market. From optimizing ad targeting to automating internal workflows, we treat AI as a tool for leverage, not just a buzzword for marketing material.
Side projects spark platform improvements. R&D isn't siloed — it’s woven into our culture. We aim high by thinking small: fast iterations, scrappy testing and a strong feedback loop between business outcomes and AI capabilities. That’s how we build smarter, faster and with fewer resources than larger players.
What are you most excited about in the field of AI right now?
Lately, I’ve been diving deep into the world of AI agents — systems designed with distinct roles, responsibilities and the ability to operate semi-independently. The idea that these agents can go beyond simple instruction-following to become collaborative, task-oriented partners is incredibly exciting. These agents can plan, prioritize and interact with tools or even each other to achieve goals with minimal human input.
What fascinates me most is the democratizing potential of this shift. Until recently, innovation at scale has often been limited to organizations with massive resources and teams. But autonomous AI agents are changing that equation. They offer leverage for small teams, solo entrepreneurs and startups, allowing them to punch well above their weight.
Imagine a two-person team managing what used to take 10 — handling customer support, internal operations, content creation, data analysis and even product research — simultaneously, thanks to a constellation of smart agents. This newfound operational efficiency doesn't just reduce costs; it creates more space for creativity, experimentation and R&D.
How do you learn from one another and collaborate?
The most successful teams I’ve worked with are built on strong relationships and open communication. Learning thrives when people feel safe to share ideas, ask questions and challenge norms — something that starts with trust.
On our team, collaborative learning often happens informally: someone drops a podcast in Slack, shares a tool they’re testing or flags a piece of code worth rethinking. These small moments spark deeper conversations and sometimes lead to major improvements.
We’ve seen side projects and experiments directly influence our roadmap — whether it’s a codebase refactor, a new ML service or a shift in architecture to support scale. The person who brings it forward often leads the implementation, which builds real ownership and confidence.
We also run regular “lunch and learn” sessions where anyone can demo a tool, break down a concept or get feedback on an idea.
