Sep 1, 2025
Design isn’t just about solving problems. It’s about choosing which problems to solve, at speed, with clarity, and in alignment with real needs. That’s where discovery matters most.
For years, design teams have operated within the rhythms of the double diamond: diverge to explore the right problem, converge to define it; diverge to explore solutions, converge to deliver them. This framework helps tame complexity. But what happens when access to real user voices becomes instant and infinite?
Continuous Discovery Meets Continuous Dialogue
In a world shaped by continuous discovery, the best teams no longer treat user research as a phase. It’s a habit. They integrate small, frequent learning moments into everyday decisions, asking questions weekly, not quarterly.
But that ambition collides with constraints:
Research bottlenecks.
Recruitment delays.
The fatigue of always waiting to ask.
AI panels challenge those constraints. They enable product teams, designers, and strategists to simulate rich user feedback in seconds. Not as a replacement for every interview, but as a complement to the messy, human reality of design.
In short: when asking becomes easy, asking becomes constant.
Recipe for Introducing AI into Discovery
Before you implement AI into your research workflow, try this first:
Imagine what happens if it works. Imagine what happens if it doesn’t.
In our collaboration with Raiffeisenbank, we designed a speculative exercise rooted in design futures thinking. The goal wasn’t to align on tools, it was to reveal hidden assumptions about how AI might transform the design process itself. What would change if anyone on the team could talk to users, instantly, forever? What tensions would emerge? What rituals would break?
This exercise isn’t about tool adoption. It’s about revealing what’s beneath the surface: organizational assumptions, research taboos, and power structures. It uses speculative design to help teams imagine the consequences (good and bad) of giving everyone access to synthetic user insight.
Start with one provocative premise:
“Imagine that starting tomorrow, anyone in this company can talk to a simulated user. Instantly, accurately, and as often as they want. No budget. No delay. No gatekeeping.”
Now ask two questions:
What’s the best that could happen?
What’s the worst?
These are the "Heaven" and "Hell" scenarios. You’re not aiming for consensus. You’re making tensions visible.
Here’s how to run it:
Frame the future. Explain the shift clearly: AI panels give us continuous, scalable access to user insight. The goal is to imagine how that could transform not just research, but the way decisions are made.
Split the room. Assign one group to map the Heaven World: breakthroughs, positive system shifts, new capabilities, better outcomes. The other group maps the Hell World: unintended consequences, failure modes, misalignments, misuses. Encourage honesty and specificity, what would actually happen in your organization?
Read and reflect. Each group reads the other’s map without editing. Ask: Which of these futures are already partially true? Which parts feel uncomfortable? What’s inevitable vs. avoidable?
Surface tensions. Bring the room together to name contradictions: empowerment vs. overload, speed vs. depth, access vs. expertise. These are the design questions that will shape AI’s role in the organization.
(Optional) Reframe as opportunities. Use the most interesting tensions to write a few speculative “How might we” questions. Let them guide deeper exploration, or spark future research strategy.
This exercise works best before rollout, when people still think AI is “a research tool.” It helps them realize: it’s actually a shift in how knowledge moves. That shift will affect who gets to ask, who gets heard, and what “research” even means.
We used this approach at Raiffeisenbank. The result: 40% more stakeholders fended up interacting with AI customers in the design sprint. Not just designers or researchers, but ops, business owners, and compliance. By imagining the futures first, they were ready to use the tools, without losing the plot. Here are the opportunities and worries we uncovered in our exercise:
🟢 Opportunities
“We can finally test things without waiting weeks.”
“Client voice throughout the sprint, not just on Day 5.”
“Designing for edge cases isn’t a luxury anymore.”
“No more post-sprint ‘let’s do a separate study’ delays.”
“We make fewer assumptions, faster.”
🔴 Worries
“What if we just start designing for the easiest answer?”
“Will we be tempted to skip hard questions if data is always available?”
“Could we stop designing for actual humans and just design for what looks good on a dashboard?”
“Are we really mapping journeys or just interrogating endpoints?”
→ Our AI Tarot Card collection can also help you imagine the future!
For a lot of design-led companies, user-centricity isn’t a buzzword, it’s an operating principle. But user-centricity changes when you can talk to simulated humans, anytime, about anything.
It means designing with more voices at the table.
It means de-risking fast decisions, not delaying them.
It means getting out of our own echo chambers—not deeper into them.
The real unlock is cultural: teams start to behave as if the user is always available. Because now, they are.
In our Research Maturity Model (inspired by the NN/g UX Maturity Model) we observe that teams who embed tools like Lakmoos evolve from reactive research toward a culture of ongoing, accessible insight. They shift from design forusers to design with user logic built into every step.
Final Thought: Ask More. Assume Less.
Designing with Lakmoos doesn’t mean replacing your users. It means bringing their logic, their context, and their messy trade-offs closer to every decision. And when every team member from designers to stakeholders has the power to ask good questions on demand, design becomes more democratic.
In a world of infinite opinions, the best design teams will be the ones who learn how to listen with precision, at scale, and continuously.
