TechVoices

VIDEO PODCAST

Auros’s Baran Erkel: Human Feedback, AI Agents, and the Future of Product Design

James Maguire

October 2, 2026

Auros Chief Strategy Officer Baran Erkel sees human feedback becoming a critical layer of the emerging agentic AI stack. Formerly known as UserTesting, Auros is expanding beyond traditional user research by connecting its network of human participants and customer insights directly into AI workflows through technologies such as Model Context Protocol (MCP).

In this TechVoices interview, Erkel discusses how enterprises can bring customer feedback into AI development loops, how humans can evaluate increasingly unpredictable AI agents, and why he believes automation will accelerate product development without eliminating the need for human judgment. As AI tools become broadly available, he argues that customer understanding — and the human insight behind it — will become an increasingly important source of competitive differentiation.

‍

Core Takeaways

Human Feedback Enters the AI Loop

Auros is turning customer research from a specialized, project-based activity into infrastructure that product managers, engineers, designers and AI agents can access directly from tools such as ChatGPT and Claude.

Agents Can Test and Refine Products

Auros is already seeing AI-forward customers use agents to design experiences, validate them with human feedback and update them, although Erkel says fully autonomous product development should not replace human judgment for major decisions.

Humans Are Needed to Evaluate AI

As AI agents become more non-deterministic and users develop more relationship-like interactions with them, traditional functional testing is insufficient. Auros is applying its human network to AI evals and to evaluating the quality of the user experience itself.

Human Judgment Creates Differentiation

Erkel argues that companies will increasingly share access to the same AI models, creating a risk that products converge toward the same average experience. Human understanding of customers, taste and strategic judgment will remain essential to creating differentiated products.

KEY QUOTES

Customer Feedback Becomes AI Infrastructure

“The vision is really to take the act of gathering customer input and feedback, customer insights, from being a project-based activity done by specialists with a special tool to becoming infrastructure that is available to anyone to use in the new ways of working.”

“Now a product manager, even an engineer, of course a designer or a user experience researcher like we've always served, can be in Claude, can be in ChatGPT where they typically work all day now, designing, building, planning, and they can kick off a study or they can interrogate past research to understand their customers better and make better decisions in these faster AI build cycles.”

How Far Can Autonomous Product Development Go?

“We have some very AI-forward customers pushing the envelope, doing things like you're talking about. And so agents designing experiences or parts of products and experiences, validating them, updating them based on that feedback, even shipping them to production. I haven't seen that exact full autonomous life cycle yet, but I'm sure that some customers will push that in the future.”

“For things like that, the smaller decisions, maybe the ones where you have more trust, the humans will allow the agents to do things like that. I don't think every decision will be autonomously driven by agents, nor should it. For the big decisions where human judgment, tastes, nuance matters, the humans will work with AIs to make those decisions.”

AI Summaries Cannot Replace Human Nuance

“AI is great at generalizing things and condensing things. We don't expect AI to be conveying the nuance in a summarized form. What we always do is we tie back to the actual human feedback. And so our users, our customers can actually see the summaries, but then always go back to the evidence and see that smile, that laugh, that confusion in the user's eyes.”

“Things like nuance, things like finding the minority of your customers or users that maybe have a different opinion than the majority which would get summarized and show up more in an AI summary, those are super valuable because that's where you can find differentiation. That's where you can find the alpha, as people say now, and maybe have new novel ideas.”

Why Humans Will Still Make the Big Product Decisions

“I think humans will be in charge of product design, the big decisions around product design. If you're not, then you're going to look like everyone else. Everybody's going to have the same models. Now you'll try to provide the context to make them different, but everybody will start to look like the median.”

“And so to differentiate, to win, to have an edge, the humans will, I strongly believe, still make the big decisions around which products to introduce, which customers and segments to go after, what the core experience should be. But then the simpler decisions that used to take longer and used to maybe require a meeting — if we know what the best practice is for a specific flow — then the agents can do that.”

ABOUT THE AUTHOR

James Maguire

Executive Director

An award-winning journalist, James has held top editorial roles in several leading technology publications, covering enterprise trends in cloud computing, AI, data analytics, cybersecurity and more. He regularly communicates with industry analysts and experts and has interviewed hundreds of technology executives.

Cloudera's Venkat Rajaji: advantages of open source AI

Snowflake's Christian Kleinerman: AI and the speed of data access

David Linthicum: LLM trends

Extreme Networks' Nabil Bukhari: how AI democratizes technology

Komprise's Krishna Subramanian: an AI model for every task

Ahead of the curve, every week.

Get in touch with us.

Ahead of the curve, 
every week.