LATERAL AI | PO THINKING | AIYING | #004
PO: Every Problem Is an Employer
From Generating More Ideas to Recruiting the Right Possibilities
What if problems could recruit ideas instead of waiting for people to discover solutions?
By Ravindra Dastikop | AIYING
1. The Provocation: Every Problem Is an Employer
Organizations are surrounded by problems.
Some are visible: rising costs, dissatisfied customers, delayed projects, and declining productivity.
Others remain hidden in routine work, outdated assumptions, and processes that nobody has questioned for years.
Meanwhile, Generative AI makes it increasingly easy to generate candidate ideas.
Yet having more ideas does not automatically mean solving more problems.
What if we reverse the familiar relationship?
Instead of people continually generating ideas and then searching for applications, imagine that problems themselves could recruit the ideas they need.
PO: Every Problem Is an Employer.
A problem cannot literally employ anyone.
It cannot advertise a vacancy, conduct interviews, or offer a salary.
That deliberate impossibility is the starting point of the provocation.
Rather than rejecting the statement as absurd, we use it to move toward new possibilities.
What would an organization look like if its problems actively recruited solutions?
2. The Fourth Step in the Lateral AI Series
This proposition follows three earlier provocations.
PO #001: GenAI Is Data-Borne
Data becomes a foundation from which AI systems can generate new possibilities.
PO #002: GenAI Generates a Possibility Dividend
The growing supply of candidate ideas creates potential value, but realizing that value requires productive opportunities.
PO #003: Ideas Need Employment
Ideas need meaningful problems, human sponsors, experiments, and opportunities to contribute.
PO #004: Every Problem Is an Employer
Now the perspective changes.
Instead of asking how to find work for ideas, we ask how problems can attract, select, and test ideas.
DATA-BORNE
Where possibilities originate
↓
POSSIBILITY DIVIDEND
Why possibilities matter
↓
IDEAS NEED EMPLOYMENT
How possibilities find productive work
↓
EVERY PROBLEM IS AN EMPLOYER
How problems recruit possibilities
Together, these provocations suggest a broader innovation architecture connecting data, possibilities, problems, experiments, and accomplishment.
3. The Hidden Assumption: Innovation Begins with Ideas
Many innovation programmes begin by inviting people to submit ideas.
Organizations run brainstorming sessions, hackathons, innovation competitions, and suggestion campaigns.
The resulting ideas are collected, evaluated, and sometimes implemented.
These activities can be valuable.
But they often assume that the most important starting point is idea generation.
Innovation begins when people generate ideas.
What if we challenge that assumption?
What if innovation begins when an organization makes its unsolved problems visible, specific, and available for exploration?
Instead of asking employees to contribute more ideas, it could invite them to help solve carefully described problems.
Instead of maintaining only an idea repository, it could maintain a living inventory of problems seeking solutions.
The shift is from a supply-led model of innovation toward a problem-led model.
4. Authentic Lateral Thinking: Provocation and Movement
Edward de Bono's lateral thinking uses deliberate provocation to escape established patterns of thinking.
A PO statement is not a claim that must be proved true.
Its purpose is to create movement toward possibilities that conventional reasoning might overlook.
Let us apply that approach.
PO: Every Problem Is an Employer.
Movement 1: Problems Advertise Vacancies
Employers advertise positions when they need work performed.
What if organizations advertised unsolved problems in the same way?
A problem vacancy might describe:
- The problem and its context
- The outcome required
- The constraints that must be respected
- The people affected
- The evidence needed to demonstrate improvement
Emerging possibility: A Problem Vacancy Board.
Movement 2: Problems Write Job Descriptions
A job description explains what a successful candidate must accomplish.
What if every important problem had an equivalent description?
For example, instead of stating:
We need innovative customer service ideas.
The organization could state:
We need to reduce repeated customer contacts without compromising service quality, privacy, or customer satisfaction.
The second statement creates a more useful opportunity for candidate ideas.
Emerging possibility: A Problem Job Description.
Movement 3: Problems Interview Ideas
Employers interview candidates to evaluate their suitability.
What if problems interviewed ideas?
A candidate idea would be asked:
- Which part of this problem can you address?
- What assumptions must hold for you to work?
- What resources do you require?
- What risks might you introduce?
- How could your contribution be tested?
Emerging possibility: An Idea–Problem Fit Interview.
Movement 4: Problems Offer Probation
Employment sometimes begins with a probationary period.
What if ideas were offered limited trials before full adoption?
An idea could receive a small experiment, prototype, simulation, or controlled pilot.
Its suitability would be judged using evidence.
Emerging possibility: A Solution Probation Programme.
Movement 5: Problems Recruit Across Boundaries
Employers may recruit people from different industries and disciplines.
What if problems could recruit ideas from unrelated fields?
A hospital scheduling problem might learn from airline operations.
A software reliability problem might learn from manufacturing quality control.
A classroom engagement problem might learn from game design.
These are analogies to explore, not ready-made solutions.
Emerging possibility: A Cross-Domain Idea Recruitment System.
The Discovery
The provocation has moved us from a conventional brainstorming exercise toward a different organizational mechanism.
What if organizations created recruitment platforms where problems attract ideas?
5. The Problem Recruitment Platform
A Problem Recruitment Platform is a proposed system that connects well-defined problems with candidate ideas, relevant expertise, responsible sponsors, and structured experiments.
The name is a deliberate extension of the employment metaphor.
It is not presented here as an established industry standard or a proven product category.
The platform would organize innovation around the demand for solutions.
How It Could Work
-
Publish the Problem
A team describes an unsolved problem and the outcome it needs. -
Define the Vacancy
The team specifies constraints, available evidence, resources, and success criteria. -
Recruit Candidate Ideas
Employees, experts, and GenAI-assisted workflows propose alternative approaches. -
Interview the Candidates
Human reviewers assess relevance, feasibility, novelty, risk, and expected value. -
Offer a Trial Assignment
Promising ideas receive small, controlled experiments. -
Hire, Adapt, or Reject
The team adopts validated ideas, modifies promising candidates, or discontinues unsuitable approaches.
PROBLEM VACANCY
↓
CANDIDATE IDEAS
↓
IDEA–PROBLEM MATCHING
↓
HUMAN EVALUATION
↓
CONTROLLED EXPERIMENT
↓
VALIDATED ACCOMPLISHMENT
The important change is not merely using AI to produce more suggestions.
It is creating a disciplined mechanism through which real problems can discover and test useful possibilities.
6. The Problem Vacancy: A New Unit of Innovation
An organization might describe a vacancy for an employee using a job description.
A Problem Recruitment Platform could describe an innovation opportunity using a Problem Vacancy Card.
Problem Vacancy Card
Problem ID: PR-001
Problem Title: Reduce Repeated Customer Contacts
Problem Owner: Customer Service Manager
Current Situation: Customers frequently contact support more than once for the same issue.
Desired Outcome: Reduce avoidable repeat contacts while maintaining service quality.
Constraints: Privacy, accessibility, accuracy, cost, and customer satisfaction.
Evidence Available: Appropriately anonymized service records and customer feedback.
Candidate Ideas Wanted: Workflow improvements, communication changes, self-service tools, and process redesign.
Trial Opportunity: Limited pilot involving a suitable customer segment.
Success Measures: Repeat contact rate, resolution quality, satisfaction, and cost.
The card does not prescribe a solution.
It creates a meaningful opportunity for different kinds of ideas to compete and collaborate.
That distinction matters.
A problem described too narrowly may recruit only familiar solutions.
A problem described too vaguely may attract suggestions that cannot be tested.
The quality of the problem description influences the quality and diversity of the possibilities explored.
7. The AI Engineering Example: Recruiting Ideas to Reduce Inference Cost
Consider an AI engineering team operating an application that uses a large language model.
The team faces increasing inference costs.
A conventional approach might be to ask:
How can we reduce LLM inference costs?
GenAI could generate a list of suggestions.
But the Problem Recruitment approach begins with a vacancy.
Problem Vacancy: AI Inference Cost Reduction
Recruit ideas that can reduce cost per successful request while maintaining acceptable response quality, latency, privacy, reliability, and user experience.
Candidate Ideas
- Cache suitable repeated responses
- Route simpler requests to smaller models
- Reduce unnecessary prompt context
- Batch compatible requests where latency permits
- Eliminate redundant model calls
- Use deterministic software for tasks that do not require an LLM
Each candidate has a different profile.
Caching may be unsuitable for highly personalized or rapidly changing information.
Model routing requires quality evaluation across request types.
Prompt reduction may remove useful context.
Batching can affect latency.
Replacing model calls with deterministic logic may work well for some tasks but not others.
The Interview
The team evaluates each candidate against:
- Cost per successful task
- Task quality
- Response latency
- Operational complexity
- Security and privacy
- Reliability under realistic workloads
The Probation
The strongest candidates receive controlled experiments using representative workloads and agreed evaluation criteria.
The team measures whether savings are real and whether unacceptable regressions occur.
The Accomplishment
Only validated improvements are adopted.
The outcome is not a longer list of optimization ideas.
It is an evidence-based engineering improvement.
PO: The inference cost problem has recruited its own solutions.
8. The Surprising Possibility: Problems as AI Agents
The metaphor suggests a more experimental possibility.
What if a problem were represented by a software agent?
Such an agent would not literally possess intentions or independent organizational authority.
It would be a software workflow acting on a problem definition supplied and governed by people.
A Problem Agent might:
- Maintain a structured description of an unsolved problem
- Identify missing information and assumptions
- Search approved knowledge sources for relevant approaches
- Generate candidate ideas with an AI model
- Suggest cross-domain analogies
- Compare ideas against predefined constraints
- Prepare experiment proposals for human review
- Track evidence from approved experiments
The agent could support problem owners, but should not independently authorize consequential deployments or bypass required oversight.
The architecture might include a problem registry, a retrieval system, a candidate generation workflow, an evaluation component, and a human approval process.
PROBLEM AGENT
Under human-defined goals and controls
↓
SEARCH + GENERATE + COMPARE
↓
HUMAN REVIEW
↓
APPROVED EXPERIMENT
↓
MEASURED OUTCOME
This is a design possibility to explore, not evidence that autonomous problem-solving agents are universally effective.
The real value would depend on data quality, evaluation methods, system reliability, costs, and human accountability.
9. The Role of Lateral Thinking in Recruitment
A recruitment platform could easily become a system that matches problems only with familiar solutions.
That would improve organization but might not produce much creative movement.
Lateral thinking introduces a different capability.
It challenges the assumptions defining the vacancy itself.
Example: A Hospital Waiting-Time Problem
A conventional problem statement might be:
How can we reduce waiting time?
A typical solution search may focus on staffing, scheduling, and queue management.
Now introduce a provocation:
PO: The Patient Never Waits.
This is not a claim that waiting can literally be eliminated.
We use it to explore movement.
What if the patient's experience began before arrival?
What if certain administrative steps were completed in advance?
What if waiting became predictable rather than uncertain?
What if some visits could be handled through appropriate remote services?
What if the organization measured unnecessary waiting separately from clinically necessary time?
These possibilities would require clinical, operational, accessibility, and safety review.
The key point is that the provocation changes which kinds of ideas the problem can recruit.
The recruitment platform supplies structure.
Lateral thinking expands the range of candidates.
10. Five Applications Beyond AI Engineering
Application 1: Education
A college publishes a problem vacancy:
Improve students' ability to apply AI engineering concepts in practical projects.
Candidate ideas may include peer learning, guided labs, case studies, short project challenges, or industry mentoring.
Faculty and students test promising approaches and assess actual learning outcomes.
Application 2: Manufacturing
A production team publishes a vacancy for ideas that can reduce avoidable equipment downtime.
AI-assisted exploration may suggest maintenance scheduling improvements, better monitoring, or workflow changes.
Engineers evaluate technical feasibility and safety before controlled trials.
Application 3: Customer Experience
A company publishes customer problems instead of broad calls for innovative ideas.
Teams propose approaches targeted at specific customer frustrations.
The strongest candidates are tested against customer outcomes.
Application 4: Public Services
A public service organization describes a problem such as reducing avoidable delays in processing applications.
Candidate ideas might involve clearer instructions, better workflow design, or appropriate digital assistance.
Any implementation would need to respect accessibility, fairness, privacy, and applicable public-sector requirements.
Application 5: Entrepreneurship
An entrepreneur begins with a portfolio of validated customer problems rather than a portfolio of product ideas.
AI helps generate alternative approaches to each problem.
Customer research and small experiments determine which possibilities merit further investment.
The emphasis moves from enthusiasm for a concept toward evidence of a problem worth solving.
11. The Lateral AI Framework: FOCUS to ACCOMPLISH
The established AIYING six-stage framework can be adapted to Problem Recruitment.
FOCUS — Publish the Vacancy
Identify a meaningful problem, its context, and the outcome required.
PROVOKE — Challenge the Vacancy
Identify assumptions that restrict how the problem is understood.
Introduce a deliberate PO statement.
MOVE — Discover Unexpected Qualifications
Explore what the provocation suggests about alternative approaches.
Look for new principles, differences, and possible applications.
GENERATE — Recruit Candidate Ideas
Use GenAI and human expertise to produce genuinely different approaches.
JUDGE — Interview for Fit
Evaluate candidates against evidence, constraints, feasibility, risk, and expected value.
ACCOMPLISH — Offer a Trial and Measure Results
Assign a responsible owner, run an appropriate experiment, and decide whether to adopt, adapt, or reject the candidate.
FOCUS → PROVOKE → MOVE → GENERATE → JUDGE → ACCOMPLISH
The framework ensures that the provocation is not reduced to an unusual slogan.
It becomes a method for discovering and testing new possibilities.
12. Reusable GenAI Prompt: The Problem Recruitment Officer
Use the following prompt to explore a real organizational challenge.
Prompt: Act as a Problem Recruitment Officer
You are an innovation facilitator combining Edward de Bono's lateral thinking principles with Generative AI.
Treat the following problem metaphorically as an employer recruiting ideas.
Problem: [Describe the problem]
Context: [Describe the organization or situation]
Desired outcome: [Define the improvement needed]
Constraints: [List cost, safety, time, quality, privacy, or other constraints]
Follow this process:
- Create a clear Problem Vacancy Card describing the problem, expected outcome, constraints, owner, and success measures.
- Identify five assumptions that may be limiting how the problem is understood.
- Generate three deliberate PO provocations that challenge those assumptions.
- Demonstrate movement from each provocation. Do not merely explain why it is unusual.
- Recruit ten genuinely different candidate ideas, including at least three cross-domain approaches.
- Conduct an Idea–Problem Fit Interview for each promising candidate.
- Select three candidates for small, low-risk experiments.
- Define the evidence required to judge each experiment.
- Identify the human sponsor, risks, dependencies, and implementation requirements.
- Recommend whether each candidate should be adopted, adapted, tested further, or rejected, based on available evidence.
Distinguish facts from assumptions. Do not fabricate evidence or claim that proposed experiments have already succeeded. Seek human approval before consequential action.
13. Student Project: Build a Problem Recruitment Platform
This provocation can become a practical AI engineering project for students and early-career developers.
Project Title
Problem Recruitment Platform — An AI-Assisted Innovation Matching System
Project Objective
Build a prototype that allows users to publish problems, generate candidate ideas, evaluate matches, and prepare small experiments.
Suggested Minimum Features
- Problem Registry: Store problem descriptions, context, constraints, and desired outcomes.
- Problem Clarifier: Use GenAI to suggest questions that make a problem more precise.
- PO Generator: Generate deliberate provocations for exploring hidden assumptions.
- Movement Explorer: Help users derive alternative directions from provocations.
- Candidate Recruiter: Generate or retrieve possible approaches.
- Fit Evaluator: Compare candidates against defined criteria.
- Experiment Planner: Create small, reviewable test plans.
- Outcome Tracker: Record human-reviewed results and decisions.
Possible Technical Architecture
USER INTERFACE
↓
PROBLEM REGISTRY
↓
LLM-ASSISTED CLARIFICATION AND IDEATION
↓
OPTIONAL RETRIEVAL / SEMANTIC SEARCH
↓
CANDIDATE EVALUATION WORKFLOW
↓
HUMAN REVIEW
↓
EXPERIMENT AND OUTCOME TRACKER
A student prototype could use Python, a lightweight web interface, a relational database, and an LLM API.
Embeddings and vector search could be added when there is a meaningful collection of existing ideas or problem descriptions to retrieve.
A multi-agent architecture is optional, not a prerequisite.
Evaluation Criteria
- Clarity of problem descriptions
- Relevance of candidate ideas
- Diversity of generated alternatives
- Quality of the proposed experiments
- Traceability from problem to recommendation
- Human oversight and error handling
- Usefulness compared with a simple brainstorming baseline
The project succeeds if it helps users move from vague problems to clearer, testable solution candidates.
It does not need to demonstrate autonomous innovation.
14. Measuring the Effectiveness of Problem Recruitment
An organization should not measure success simply by the number of problems posted or ideas generated.
More useful measures may include:
| Measure | Purpose |
|---|---|
| Problem clarity | Whether vacancies specify meaningful outcomes and constraints |
| Candidate relevance | Whether proposed ideas address the actual problem |
| Idea diversity | Whether candidates represent substantively different approaches |
| Time to experiment | How quickly promising ideas receive appropriate testing |
| Validated adoption | Whether experiments lead to evidence-supported improvements |
| Net value realized | Benefits after accounting for relevant costs and risks |
These are proposed evaluation dimensions rather than established benchmarks.
The most important question is whether the approach produces better decisions and outcomes than the organization's existing innovation process.
15. Risks and Limitations
Poorly Defined Problems
A vague or misleading problem description may attract irrelevant solutions.
Premature Solution Bias
A vacancy may unintentionally prescribe a favored solution before alternatives are explored.
AI Hallucinations
GenAI may generate plausible but incorrect claims about feasibility, evidence, or expected benefits.
False Diversity
A long list of candidate ideas may contain only superficial variations of the same approach.
Evaluation Bias
Human and automated reviewers may favor familiar ideas and overlook unconventional candidates.
Implementation Risk
Promising ideas may fail when deployed in real environments.
Privacy, Security, and Accountability
Problem descriptions may contain sensitive business, personal, or operational information.
Appropriate access controls, data handling, review, and accountability are essential.
The recruitment metaphor should never imply that an AI system has independent authority to make consequential organizational decisions.
16. Five Questions for Executive Leadership
-
Do we know which problems most urgently need solutions?
Is there a visible, prioritized inventory of meaningful unsolved problems?
-
Are our innovation programmes driven by problems or by idea volume?
Do we reward the generation of suggestions or the resolution of valuable challenges?
-
Can problems recruit ideas across organizational boundaries?
Are teams able to discover relevant approaches outside their own departments?
-
Can GenAI help us discover unexpected matches?
Are we using AI only to brainstorm, or also to connect problems with diverse candidate solutions?
-
Do we provide opportunities for responsible experimentation?
Can promising ideas move quickly into small, measurable, human-supervised tests?
17. The New Human Work: Becoming a Problem Recruiter
The growing capability of GenAI changes the economics of generating candidate ideas.
But valuable innovation still requires people who can recognize meaningful problems.
It requires people who can frame challenges without prematurely narrowing the solution space.
It requires people who can recognize unconventional possibilities, evaluate evidence, manage risks, and coordinate implementation.
The Problem Recruitment metaphor suggests a useful description of emerging human work:
The human role is not only to generate ideas, but to create the conditions in which the right ideas can solve the right problems.
This does not eliminate the importance of human creativity.
It expands the range of activities through which creativity creates value.
Problem discovery, problem framing, lateral exploration, judgment, and accomplishment become central capabilities.
18. The Next Provocation Emerges
The recruitment metaphor leads to another unexpected question.
If problems are employers, what happens when a problem cannot find a suitable idea?
An employer might change the job description.
It might develop candidates.
It might redesign the work.
Or it might discover that the position itself is unnecessary.
This leads to a new provocation:
PO: Problems Need Career Counselling.
What if some problems are difficult to solve because they are framed incorrectly?
What if the most useful intervention is not a better solution, but a better understanding of the problem?
That possibility opens the next stage of lateral exploration.
19. Conclusion: Let Problems Recruit Possibilities
Generative AI has expanded our ability to generate candidate solutions.
But idea abundance does not automatically create innovation.
Organizations still need mechanisms that connect possibilities with meaningful problems.
The provocation PO: Every Problem Is an Employer invites us to reverse the conventional innovation process.
Instead of starting only with idea generation, we can make unsolved problems visible, define their needs, recruit diverse approaches, and evaluate candidates through responsible experimentation.
The proposed Problem Recruitment Platform offers one way to explore that shift.
PO: Every Problem Is an Employer.
Problems create demand.
GenAI expands the candidate pool.
Lateral thinking discovers unconventional matches.
Human judgment selects and tests.
Accomplishment creates value.
Stop asking only for more ideas. Start helping problems recruit possibilities.
20. Your Turn: Publish a Problem Vacancy
Choose one meaningful problem from your organization, classroom, business, or community.
Write a Problem Vacancy Card.
Describe the outcome required without prescribing a solution.
Use GenAI to recruit ten diverse candidate ideas.
Apply one deliberate PO provocation to challenge the original problem framing.
Select one promising candidate for a small, appropriate experiment.
Record what you discover.
The goal is not to prove that AI can generate answers.
It is to discover whether a problem-led approach produces more useful possibilities and better outcomes.
What problem in your organization would you advertise first?
Further Reading and Conceptual Foundations
The Problem Recruitment Platform is an original exploratory design concept developed through the PO metaphor in this article. It is not presented as an established or empirically validated innovation framework.
- Edward de Bono, Serious Creativity (1992) — provocation, movement, and deliberate creative thinking.
- ISO 56002:2019 — Innovation management system guidance. Explore
- NIST AI Risk Management Framework — responsible management of AI risks. Explore
- UNFPA — Demographic dividend, the economic analogy explored earlier in this series. Explore
