What If Customers Designed the Problems?
One Strange Question. A New Possibility.
What if the greatest business innovation begins not with a better solution, but with a completely different way of defining the problem?
The Story: The Product Nobody Wanted
Six months.
That was how long an ambitious Indian startup had spent developing its new AI-powered productivity application.
The company had a talented engineering team, enthusiastic founders and a clear product roadmap.
The application could organize tasks, summarize meetings, generate schedules, prioritize activities and recommend ways to improve productivity.
The team was proud of what it had built.
The interface was elegant. The technology worked. Early demonstrations received encouraging responses.
Then came the launch.
People signed up.
Some explored the application.
A few praised its features.
But something unexpected happened.
Very few customers returned.
The founders were puzzled.
The product worked exactly as designed.
So why weren't people using it?
The meeting that changed everything
The team assembled to discuss the problem.
The product manager suggested adding more features.
The marketing manager proposed a promotional campaign.
The designer recommended simplifying the interface.
The founder wanted a better onboarding experience.
Each suggestion sounded reasonable.
And each was based on the same assumption:
We have built the right solution. We just need to make customers appreciate it.
Then Nisha, a young designer who had joined the company recently, asked a question.
“Why are we trying to improve something customers may not actually need?”
The room fell silent.
She continued.
“What if we stopped deciding what problems customers should have?”
Someone laughed.
“Customers don't design products. We do.”
Nisha smiled.
“Exactly. That's what we assume.”
Then she introduced an unusual proposition:
“What if customers designed the problems, and companies competed to solve them?”
At first, the idea seemed absurd.
Customers were supposed to buy products, not define a company's innovation agenda.
But as the team explored the idea, something began to change.
They stopped discussing features.
They started discussing customers' lives.
What were people struggling with before discovering the application?
What were they already doing to manage their work?
Which frustrations were important enough to motivate a change?
What problems had the company never thought to investigate?
For the first time, the team was not asking how to sell its product.
It was asking whether it had understood the problem.
And that was the beginning of a very different conversation.
1. The Hidden Assumption Behind Product Development
Many organizations follow a familiar sequence:
Identify an opportunity → Define a problem → Design a solution → Build a product → Ask customers for feedback.
There is nothing inherently wrong with this process.
Successful organizations already use customer research, co-design, observation and experimentation to understand real needs.
Yet even customer-focused teams can become attached to their initial interpretation of a problem.
Once a solution takes shape, subsequent discussions often revolve around improving it.
Should we add another feature?
Should we change the pricing?
Should we redesign the interface?
Should we market the product differently?
These questions may be useful.
But they can overlook a more fundamental possibility:
What if we have defined the wrong problem?
Consider the difference between these two questions.
Conventional question:
“How can we make our productivity application more attractive?”
Alternative question:
“What difficulties do people actually experience when trying to accomplish meaningful work?”
The first assumes the application is the appropriate starting point.
The second leaves open the possibility that customers need something entirely different.
Perhaps their biggest difficulty is not task management.
Perhaps it is unclear responsibilities, constant interruptions, unrealistic expectations, conflicting priorities or decisions delayed by other people.
A better task-management application may not address any of those problems.
This is where deliberate lateral thinking becomes valuable.
2. Enter PO: A Provocation That Reverses the Relationship
Edward de Bono developed lateral-thinking methods to help people move beyond established patterns of thought.
One of his important contributions was the use of PO, a signal that introduces a deliberate provocation.
A PO statement is not necessarily a fact, a recommendation or a proposal to be implemented literally.
It is a thinking device.
It temporarily suspends ordinary judgment so that an unusual idea can become a stepping stone toward new possibilities.
For this episode, we use reversal.
Our PO statement
PO: Customers design the problems. Companies compete to solve them.
This reverses a familiar relationship.
Instead of companies defining problems and presenting solutions to customers, imagine customers having the primary role in framing the challenges organizations attempt to solve.
The idea may initially seem impractical.
Customers do not always know the technical possibilities.
They may disagree about what matters.
Their stated preferences may differ from their actual behavior.
And organizations still need expertise, judgment and commercial discipline.
But those objections are not reasons to abandon the provocation immediately.
They are reasons to explore it carefully.
The purpose of PO is not to prove that customers should literally control every aspect of product development.
The purpose is to discover what becomes possible when we reverse our usual way of thinking.
3. PO–AI Thinking Lab: From Provocation to Practical Innovation
Beginning with Episode #004, PO Prompts follows a refined six-stage Thinking Lab.
The approach draws on Edward de Bono's provocation and movement concepts, then extends the exploration using generative AI, critical evaluation and practical experimentation.
The six stages are:
FOCUS → PROVOKE → MOVE → GENERATE → JUDGE → ACCOMPLISH
Let's apply each stage to our startup.
Stage 1 — FOCUS: Identify the Real Problem
The startup initially described its problem as:
“Our productivity application is not retaining enough customers.”
That is a business symptom.
It does not yet explain why customers leave.
The team assumes that improving the application will improve retention.
But what if the product itself is not addressing a sufficiently important customer need?
Questions worth asking
What problem did we believe customers had?
What evidence supported that belief?
How did customers actually describe their difficulties?
What were they doing before our product existed?
Which assumptions did we make without testing?
Are we improving a solution before validating the problem?
The team discovers several untested assumptions:
Assumption 1: People want another productivity application.
Assumption 2: More features will make the application more valuable.
Assumption 3: Customers understand and describe their needs in the same language as product developers.
Assumption 4: Low retention primarily reflects problems with the product interface or marketing.
Assumption 5: The company should define the problem before involving customers.
These assumptions may or may not be true.
The point is to make them visible.
The first breakthrough
The team's challenge changes from:
“How do we improve our application?”
to:
“What problem are customers actually trying to solve, and how do we know?”
This shift creates room for the provocation.
Stage 2 — PROVOKE: Reverse the Conventional Relationship
Now introduce the PO statement:
PO: Customers design the problems. Companies compete to solve them.
Imagine a world in which companies cannot begin by proposing products.
Instead, customers publish descriptions of difficulties they experience.
Organizations study those difficulties and compete to demonstrate which solutions produce the best outcomes.
A customer might say:
“I lose two hours every day because I don't know which tasks genuinely matter.”
Another might say:
“My work is interrupted by constant requests from colleagues.”
A third might say:
“I have too many tools, and none of them helps me make difficult decisions.”
These statements describe different problems.
Yet the startup's original application might have treated all three as variations of task management.
The reversal forces the company to reconsider its starting point.
Why reversal matters
Reversal interrupts a familiar direction of thinking.
Instead of asking customers to react to a predefined solution, we imagine customers setting the problem agenda.
This does not mean that every customer statement should be accepted literally.
Customers may describe symptoms rather than underlying causes.
Organizations still need to investigate, interpret and validate what they hear.
But the reversal exposes an important principle:
Problem discovery deserves as much creativity as solution development.
Stage 3 — MOVE: Turn the Provocation into New Directions
This is a crucial part of deliberate lateral thinking.
A provocative idea should not remain merely strange or entertaining.
We must move from it toward potentially useful concepts.
De Bono's movement approaches include extracting principles, examining positive aspects, considering consequences, noticing differences and identifying circumstances where an idea might have value.
Let's use several of these approaches.
Movement A — Extract a principle
From the provocation, we can extract this principle:
The people experiencing a problem should have a meaningful role in defining what needs to change.
That principle does not require customers to become professional designers.
It suggests that organizations should create better ways to discover and interpret lived experiences.
A possible direction emerges:
Build the problem-discovery process before building the product roadmap.
Movement B — Examine positive aspects
What might be valuable about customers defining problems?
Organizations could discover frustrations they had not anticipated.
Teams could avoid building features that address weak or imaginary needs.
Customers could feel that their experiences are taken seriously.
Innovation might shift from adding functionality toward improving real outcomes.
These are hypotheses, not guaranteed benefits.
But they point toward experiments worth conducting.
Movement C — Explore consequences
Imagine the startup genuinely adopting this reversal.
What would happen next?
Its product team might spend less time debating features internally.
Its customer interviews might focus more on experiences and workarounds.
Its roadmap might become organized around verified customer difficulties rather than proposed features.
Its success measures might shift from the number of features delivered toward improvements in customers' actual work.
The team might even discover that the most valuable solution is not another application.
It could be a service, a workflow change, an integration or a simpler way of coordinating people.
Movement D — Find special circumstances
Where might customer-defined problems be particularly valuable?
Consider:
Healthcare services, where patient experiences can reveal practical difficulties that standard processes overlook.
Education, where students may identify barriers that curriculum designers underestimate.
Public services, where citizens encounter frustrations that internal administrative metrics fail to capture.
Small businesses, where owners may struggle with problems poorly represented by existing software.
Enterprise technology, where everyday users experience workflow obstacles that senior decision-makers rarely see.
These are areas for exploration, not proof that the approach will always succeed.
The movement insight
We began with an unusual statement:
“Customers design the problems.”
We have moved toward a more practical idea:
Organizations can innovate by giving customers a stronger role in revealing, framing and validating problems before solutions are developed.
That is a meaningful direction for further investigation.
Stage 4 — GENERATE: Use AI to Expand the Possibilities
Now generative AI becomes a thinking partner.
Instead of asking AI for the best product features, we ask it to expand the directions revealed by PO.
Here are five possibilities.
Possibility 1 — A Customer Problem Library
Imagine an organization maintaining a structured collection of customer-described difficulties.
Instead of recording only feature requests, it captures:
What the customer was trying to accomplish
What went wrong
What workaround they used
What consequences followed
What outcome they hoped to achieve
AI could help organize these descriptions into themes.
The organization could then investigate which problems recur, which are most consequential and which are poorly addressed.
Potential value: A stronger connection between product decisions and actual customer experiences.
Important caution: AI-generated clusters are interpretations. They require human review and validation against the underlying evidence.
Possibility 2 — A Problem-First Innovation Marketplace
Imagine customers and communities publishing challenges rather than shopping for predefined solutions.
Organizations, researchers and independent innovators could propose competing approaches.
A challenge might be:
“How can small retailers reduce the time spent reconciling daily transactions?”
Different providers might propose software, process changes, training or automated integrations.
Solutions would be judged against agreed outcomes.
Potential value: Competition shifts toward solving meaningful problems rather than showcasing impressive features.
Possibility 3 — AI-Assisted Discovery of Hidden Frustrations
Customers may not always express their needs as clear problem statements.
They might say:
“This takes too long.”
“I keep forgetting.”
“I don't trust the result.”
“I have to ask three people before anything happens.”
AI could help researchers organize interviews, compare descriptions and generate hypotheses about underlying difficulties.
For example, repeated complaints about forgetting tasks might point toward unclear priorities, excessive workload or fragmented communication.
Potential value: More systematic exploration of what customers mean, not merely what they say.
Important caution: AI cannot reliably determine an unspoken need from a few comments. Every hypothesis must be tested with real users.
Possibility 4 — Customer-Defined Success Measures
Companies often evaluate products using internal metrics such as feature adoption, session duration or engagement.
What if customers helped define what successful outcomes look like?
A customer might care about:
Completing essential work before leaving the office
Making fewer avoidable mistakes
Reducing unnecessary interruptions
Spending less time coordinating routine activities
Feeling more confident about important decisions
AI could help translate these outcome descriptions into candidate measures.
Potential value: Product success becomes more closely connected to meaningful customer results.
Possibility 5 — Products That Evolve Around Changing Problems
Imagine a product that regularly asks users:
“What has become more difficult for you since we last spoke?”
AI could help organize responses, identify emerging patterns and suggest new questions for researchers.
The organization could investigate whether changing customer circumstances require different services, workflows or features.
Potential value: Continuous problem discovery rather than occasional feature-request collection.
Important caution: Customers must understand what information is collected, how it is used and what choices they have. Privacy and consent are essential.
Stage 5 — JUDGE: Which Possibilities Are Worth Pursuing?
Creative thinking should expand possibilities before narrowing them.
But eventually, judgment is necessary.
Not every unusual idea deserves implementation.
Not every AI-generated suggestion reflects a genuine customer need.
And not every customer request represents a commercially viable opportunity.
Let's evaluate our five possibilities.
| Possibility | Potential value | Main challenge | First validation step |
|---|
| Customer Problem Library | Better visibility into recurring needs | Incomplete or biased feedback | Review a small sample of actual customer experiences |
| Problem-First Marketplace | Diverse competing solutions | Attracting participants and defining success | Test one challenge with a small group |
| AI-Assisted Frustration Discovery | Faster organization of qualitative research | Misinterpretation and hallucination | Compare AI themes with original interviews |
| Customer-Defined Success Measures | More meaningful outcomes | Selecting reliable measures | Ask customers to rank desired outcomes |
| Products That Evolve Around Problems | Earlier discovery of changing needs | Privacy and feedback fatigue | Pilot periodic opt-in feedback |
These are qualitative judgments for exploration, not evidence-based rankings.
Questions for responsible evaluation
Before selecting a possibility, ask:
Usefulness: Does this address a meaningful customer problem?
Feasibility: Can we test it with available time and resources?
Evidence: What observations support the underlying assumptions?
Risk: Could the approach create privacy, fairness, accessibility or trust problems?
Measurement: What would demonstrate that the approach works?
Alternatives: Is there a simpler or more effective way to achieve the same outcome?
For our startup, the Customer Problem Library may be a sensible starting point.
It is relatively easy to pilot and does not require building a new platform.
The team can begin with real conversations.
Stage 6 — ACCOMPLISH: A Seven-Day Practical Experiment
Let's turn the idea into an experiment.
Experiment: Stop selling the solution. Discover the problem.
Objective: Determine whether direct customer problem discovery reveals important needs that the startup's current product roadmap overlooks.
Duration: Seven days.
Participants: Five to ten customers or prospective customers, selected to represent different relevant usage experiences.
Day 1 — Define the learning goal
Write down the problem the company believes it solves.
For example:
“Busy professionals struggle to manage their daily tasks.”
List the evidence supporting that statement.
Identify what remains uncertain.
Days 2–3 — Interview customers without pitching the product
Ask open-ended questions:
Tell me about a recent workday that felt unnecessarily difficult.
What were you trying to accomplish?
What got in your way?
What did you do to work around the difficulty?
How often does this happen?
What happens when the problem remains unresolved?
Have you tried anything to improve the situation?
What would a meaningful improvement look like?
Avoid leading questions such as:
“Would you like an AI app that organizes your tasks?”
That question introduces a solution before understanding the problem.
Day 4 — Organize the evidence with AI
With appropriate consent and privacy safeguards, provide anonymized interview notes to ChatGPT.
Ask it to distinguish:
Require it to link each theme back to the relevant evidence.
Do not ask AI to invent customer needs.
Day 5 — Compare discoveries with the existing roadmap
Which problems were already known?
Which were underestimated?
Which are not addressed by the current product?
Which assumptions appear unsupported?
Select one surprising problem for further investigation.
Day 6 — Design a small solution experiment
Do not immediately build a major new feature.
Try a lightweight intervention.
For example, if customers struggle with unclear priorities, test a simple daily priority-setting workflow with a few participants.
It might involve a checklist, a short guided conversation or a manual service.
Day 7 — Evaluate and learn
Compare the experimental approach with customers' existing methods.
Ask:
Did participants experience an improvement?
Was the problem important enough for them to change behavior?
What evidence supports the result?
What remains uncertain?
Should the team continue, revise or abandon the idea?
What counts as success?
The experiment succeeds as a learning exercise if it produces a clearer, evidence-supported understanding of an important customer problem.
It does not need to produce a new product immediately.
Sometimes the most valuable accomplishment is discovering that you should not build what you originally planned.
4. The Complete ChatGPT PO Prompt
The following prompt turns this episode into a reusable AI-assisted thinking exercise.
You can apply it to a product, service, customer experience, educational program, organizational process or business model.
Copy and use this prompt
Act as my PO lateral-thinking partner, drawing on Edward de Bono's provocation and movement methods.
My situation: I am developing or improving a product, service, process or experience.
Use this provocation:
“PO: Customers design the problems. Organizations compete to solve them.”
Guide me through six stages:
1. FOCUS: Identify the conventional problem definition and hidden assumptions about who defines customer needs.
2. PROVOKE: Explore the reversal seriously without immediately dismissing it as impractical.
3. MOVE: Use principle extraction, consequences and positive aspects to discover new directions.
4. GENERATE: Develop at least five genuinely different possibilities for discovering and responding to customer-defined problems.
5. JUDGE: Evaluate each possibility for usefulness, feasibility, risks and evidence needed. Clearly separate assumptions from verified customer insights.
6. ACCOMPLISH: Help me design one small, low-cost experiment with a clear success measure.
Ask me questions at each stage. Do not jump straight to solutions.
Do not invent customer evidence, research findings or validation results.
Begin by asking:
“What product, service or customer experience would you like to rethink?”
How to use it effectively
Begin with a real challenge.
For example:
“Our training program has low completion rates.”
“Our customers rarely use our mobile application.”
“Our students are not engaging with online lessons.”
“Our consulting clients often request major changes after reviewing proposals.”
Then allow AI to guide you through the six stages.
At each stage, contribute your own knowledge, experiences and evidence.
Treat AI-generated ideas as possibilities for investigation, not as established facts.
The objective is not to produce a longer list of solutions.
It is to discover a different and potentially more useful way of understanding the problem.
5. Beyond Startups: Where Else Could This PO Prompt Be Useful?
The same provocation can open possibilities in several domains.
Education
Conventional approach: Teachers design learning activities and students complete them.
PO reversal: Students identify the learning difficulties; educators design ways to address them.
This could inspire student-led problem discovery, personalized support and new approaches to curriculum improvement.
Healthcare
Conventional approach: Healthcare organizations design services around clinical and operational requirements.
PO reversal: Patients describe the difficulties of navigating care; organizations investigate how to improve the experience.
This might reveal opportunities in communication, scheduling, accessibility and care coordination.
Clinical safety and professional expertise remain essential.
Government and Public Services
Conventional approach: Agencies define services and citizens navigate them.
PO reversal: Citizens identify recurring service difficulties; agencies redesign processes around validated needs.
This might reveal problems that administrative performance indicators do not fully capture.
Consulting
Conventional approach: Consultants receive a client's stated problem and propose a solution.
PO reversal: Clients and affected stakeholders collaboratively redefine the problem before recommendations are developed.
This could reduce the risk of producing sophisticated solutions to poorly understood challenges.
AI Product Development
Conventional approach: Developers build AI capabilities and search for applications.
PO reversal: Users identify consequential problems; developers determine whether AI is appropriate.
Sometimes the best solution will involve AI.
Sometimes it will involve a simpler process, clearer responsibilities or better information.
The problem should guide the technology—not the other way around.
6. Why PO and Generative AI Work Well Together
Generative AI has an extraordinary capacity to produce ideas, variations, scenarios and alternative formulations.
But generating more ideas is not necessarily the same as thinking differently.
If we ask AI to improve an existing product, it may produce numerous plausible improvements while accepting the product's underlying assumptions.
PO introduces a deliberate disruption.
It changes the direction of exploration.
AI can then help develop the implications of that disruption.
Human judgment determines which possibilities are worth investigating.
Experimentation provides evidence about what actually works.
The PO–AI partnership
PO provokes.
It challenges assumptions and disrupts established thinking patterns.
AI generates.
It helps explore alternatives, connect concepts and develop hypothetical possibilities.
Humans judge.
They bring context, values, responsibility and critical evaluation.
Together, they accomplish.
They turn promising ideas into experiments, decisions and practical improvements.
This is the philosophy behind the AIYING adaptation of PO thinking.
It is not a replacement for Edward de Bono's original methods.
It is an attempt to apply deliberate creative thinking in an age of generative AI.
7. The Bigger Insight: Who Owns the Problem?
Organizations often compete through the quality of their solutions.
Better technology.
More features.
Lower prices.
Faster delivery.
Improved customer experience.
All of these can matter.
But what if a deeper competitive advantage comes from understanding problems that others have overlooked?
Consider two organizations.
Organization A asks:
“How can we make our solution better?”
Organization B asks:
“What important problem are we failing to understand?”
Organization A may improve an existing offering.
Organization B may discover an entirely new opportunity.
Neither question is universally superior.
But when an organization becomes too attached to its existing solution, the second question may be especially valuable.
The PO reversal helps make that question visible.
Innovation does not always begin with a brilliant answer.
Sometimes it begins by giving someone else the opportunity to redefine the question.
8. Your PO Challenge
Think about something you are currently trying to improve.
It could be a product, service, course, business process, family routine or community initiative.
Ask yourself:
Who currently defines the problem?
Then introduce the provocation:
PO: The people affected by the problem become its designers.
What changes?
What hidden assumptions become visible?
What possibilities emerge?
What could you test this week?
Try the complete ChatGPT prompt above and share the most surprising direction you discover.
A question for readers
If your customers, students, employees or users could redefine the problem you are trying to solve, what might they tell you that you have never considered?
I'd be interested in hearing your examples.
About PO Prompts
PO Prompts is an AIYING series exploring how deliberate provocations can help us discover new possibilities with generative AI.
Each episode combines:
One relatable story
One hidden assumption
One deliberate PO provocation
Structured movement toward new possibilities
AI-assisted exploration
Human judgment and evaluation
One practical experiment
One complete, reusable ChatGPT prompt
The series draws inspiration from Edward de Bono's work on lateral thinking, provocation and movement.
The six-stage PO–AI Thinking Lab is an AIYING adaptation that combines these ideas with generative AI and practical experimentation.
One Image. One Story. One Provocation. Many Possibilities.
PO Prompts — One Strange Question. A New Possibility.
AIYING — Accomplishing with AI.
Further Reading
Edward de Bono, Serious Creativity — an introduction to deliberate creativity, provocation and movement.
https://www.debono.com/serious-creativity-article
The startup story and characters in this article are fictional and used to illustrate the thinking method. The suggested applications and experiments are exploratory, not reported research findings.
