As AI agents become more capable, they are no longer just tools; they are becoming active participants in the product development process. The key question for product teams is: how do we design products for a world where AI agents, not just humans, are part of the user journey?
To explore this topic, we brought together three expert speakers:
Mehdi Boudoukhane, co-founder & CEO of Cycle, a product feedback platform leveraging AI to help teams extract insights from customer feedback.
Ayan Barua, co-founder & CEO of Ampersand, a developer platform for product integrations.
Timoté Geimer, CEO of dualoop, product management consultancy working with companies across Europe to shape products and teams.
AI won’t replace you, but someone using AI will
AI is most powerful when it enhances human capabilities rather than replacing them. The best product teams use AI to improve decision-making, research, and execution speed.
AI can help product managers refine their work before handing it off to engineers. Instead of sending a spec as-is, a PM can ask AI, “How would a skeptical engineer critique this?” or “What might a CEO find unclear?” This kind of iterative feedback loop helps teams preempt issues and improve communication.
But there’s a catch: AI can’t think for you. Without strong product intuition, AI tools can lead to misinformed decisions and overconfidence in flawed outputs. AI is only useful if the person using it knows how to ask the right questions and critically assess the results.
The bullsh*t AI test: does this feature really matter?
Many companies rush to integrate AI without questioning whether it adds real value. To filter out unnecessary AI features, ask three key questions:
- Would this product still be useful without AI?
AI should enhance a core function, not be the entire value proposition. If the product has no clear purpose without AI, it’s likely built on hype.
- Does AI make the core experience at least 10x better?
AI should significantly amplify the value of a product, not just make it slightly more efficient. A feature that improves speed by 20% is useful, but it’s not transformative.
- Could a user accomplish this with ChatGPT and a well-crafted prompt?
If users can replicate your AI feature by copy-pasting text into a free AI tool, it won’t be a long-term differentiator. The best AI features leverage proprietary data, deep user context, and direct integrations to offer something unique.
Many AI-driven products today fail this test. A simple wrapper around OpenAI’s API isn’t a product, it’s a temporary hack that will become obsolete with the next model release.
The UX shift: designing for AI agents, not just humans
For decades, UI/UX has been built around human interaction. Buttons, forms, and dashboards help users navigate complex systems. But AI agents are new types of users, and they don’t need traditional interfaces.
- AI agents interact with APIs, not buttons. Instead of navigating a UI, they consume data directly and execute tasks.
- Authentication and permissions need to evolve. If an AI agent is operating on behalf of a human, how do we control what it can access and modify?
- AI-driven automation is reducing manual work. If an AI agent handles 80% of a task, what role does UI/UX play in guiding the remaining 20%?
This shift means products need to be designed differently. Instead of thinking about UI first, teams should consider how humans and AI will collaborate, and where interfaces are actually necessary.
What’s next for AI in product?
The AI revolution is still in its early stages, but a few trends are becoming clear:
- Faster time to value: AI-powered tools are cutting the time between sign-up and first success. Products like Gamma and V0 let users describe what they need and generate a first version instantly.
- Smaller, more efficient teams: AI is allowing startups to scale with fewer people. Some of today’s fastest-growing AI startups are reaching millions in ARR with teams of 20 people or less.
- AI as a collaborator, not just an automation tool: The future isn’t about replacing jobs, it’s about extending human capabilities. AI will become a real-time thought partner that helps people make better decisions, faster.
Final thoughts
AI isn’t a magic bullet. The best products will integrate AI in ways that genuinely enhance the user experience, reduce friction, and unlock new possibilities. AI should be a means to an end, not the end itself.
As AI becomes standard in every product, real differentiation will come from those who use it thoughtfully and strategically. The question isn’t how do we add AI?, it’s how do we use AI to solve real problems in better ways?
