LATERAL AI | AIYING | STRATEGIC THINKING
PO: GenAI Is Data-Borne
From Data-Driven Decisions to Data-Borne Possibilities
A lateral thinking exploration of how Generative AI changes the relationship between data, intelligence, imagination, innovation, and human work.
By Ravindra Dastikop | AIYING
1. The Provocation: GenAI Is Data-Borne
For decades, business and technology leaders have been encouraged to become data-driven.
Organizations invested in databases, enterprise applications, data warehouses, business intelligence, analytics, and machine learning.
The central promise was straightforward:
Better data leads to better information. Better information leads to better decisions. Better decisions lead to better outcomes.
This logic continues to be valuable.
But Generative AI invites us to explore a different relationship between data and intelligence.
PO: GenAI Is Data-Borne.
What if data is not merely something intelligence consumes, but part of the foundation from which new forms of machine-generated possibilities emerge?
This is deliberately a provocation rather than a formal scientific definition.
Generative AI systems depend on much more than data: model architectures, algorithms, computational resources, training objectives, evaluation, and human engineering are also essential.
Yet the expression data-borne offers a productive new way to think about what changes when data becomes part of a generative system.
The important question is no longer only:
What can our data tell us?
It is also:
What can our data help us imagine, create, and accomplish?
2. From Data-Driven to Data-Borne
Consider three different ways of describing the relationship between data and intelligent systems.
| Perspective | Primary Focus | Typical Question |
|---|---|---|
| Data-driven | Evidence and decisions | What does the data reveal? |
| Data-trained | Learning statistical patterns | What patterns can a model learn? |
| Data-borne | Generating possibilities | What new outputs and alternatives can emerge? |
These are not three mutually exclusive technical categories or necessarily successive stages of AI development.
They are three complementary perspectives.
A single generative AI system can be data-trained, support data-driven decisions, and generate new possibilities.
The value of the third perspective lies in changing what we pay attention to.
3. The Lateral Thinking Connection
Edward de Bono developed lateral thinking as a deliberate approach to escaping established patterns of thought and generating alternatives.
One of its distinctive devices is the provocation, often signalled by the word PO.
A provocation is not necessarily a statement we believe to be true.
Its purpose is to move thinking away from familiar assumptions.
Consider the familiar assumption:
Data exists to describe reality and support decisions.
Now introduce the provocation:
PO: Data exists to generate possibilities.
The provocation does not invalidate the original purpose. Instead, it creates another direction for exploration.
Movement from the Provocation
Rather than immediately judging whether the statement is correct, we explore where it takes us.
- What if a database became a source of product ideas?
- What if customer complaints became innovation prompts?
- What if organizational history became a simulation resource?
- What if reports generated questions instead of only answers?
- What if the value of data were measured partly by the alternatives it helped people discover?
This is the essential movement:
Data as Record → Data as Pattern → Data as Possibility
The progression is conceptual, not a claim that generative models automatically produce reliable innovations.
4. Data Was Traditionally About the Past
Most enterprise data systems began by recording events.
A transaction occurred. A customer purchased a product. A machine failed. An employee completed a task. A student received a grade.
The system stored evidence of something that had happened.
Analytics then helped organizations identify patterns, compare performance, forecast outcomes, and make decisions.
Machine learning extended these capabilities through prediction, classification, anomaly detection, and other forms of statistical inference.
Generative AI adds another capability: it can produce novel combinations and representations based on learned patterns and supplied context.
For example, an AI assistant working with authorized enterprise information might generate:
- Alternative product concepts
- Customer communication drafts
- Possible process redesigns
- Training simulations
- Software prototypes
- Competing strategic scenarios
These outputs are possibilities, not verified truths.
They must be evaluated against evidence, constraints, feasibility, and real-world outcomes.
5. A New Strategic Question: What Is the Value of Data?
Traditionally, organizations have valued data for its usefulness in reporting, compliance, operational efficiency, prediction, and decision-making.
These remain important.
But the data-borne perspective introduces an additional dimension:
The value of data may also lie in the range of useful possibilities it helps people discover and test.
Imagine two organizations possessing similar customer service records.
Organization A uses those records to measure complaint volumes, response times, and satisfaction.
Organization B does the same, but also uses appropriately governed AI systems to explore alternative service models, identify unmet needs, and prototype new customer experiences.
Organization B is not automatically more successful. Its advantage depends on the quality of its ideas, execution, governance, and market validation.
Nevertheless, it has expanded the questions its data can help answer.
That expansion is strategically significant.
6. The Content Dividend and the Possibility Dividend
Generative AI has dramatically reduced the effort required to produce many forms of digital content.
Text, images, software, presentations, learning materials, and synthetic examples can now be generated rapidly.
This suggests what we might call a content dividend: the potential increase in productive content creation made possible by generative technologies.
But abundant content creates another challenge.
More content does not automatically produce more attention, trust, understanding, or value.
The same principle applies to ideas.
GenAI can produce hundreds of alternatives. That does not mean those alternatives are original, feasible, desirable, or valuable.
We therefore need to distinguish between:
- Content abundance: More material can be produced.
- Possibility abundance: More alternatives can be explored.
- Value creation: Selected alternatives solve meaningful problems.
The opportunity is not simply to generate more.
It is to turn expanded possibilities into better accomplishments.
7. Human Imagination Becomes More Important
If GenAI can generate content and alternatives at increasing speed, what remains distinctive about human contribution?
A useful answer is not that humans must compete with machines in producing more outputs.
It is that humans must become better at deciding which possibilities matter.
Consider five human responsibilities.
Purpose
Why are we exploring this problem? What outcome is worth pursuing?
Framing
Are we asking the right question? Are hidden assumptions limiting the possibilities?
Imagination
What unconventional directions should we explore, including those not suggested by existing patterns?
Judgment
Which ideas are credible, ethical, feasible, and valuable?
Accountability
Who takes responsibility for the consequences of deploying an AI-generated solution?
These responsibilities are not exclusive technical capabilities of humans in every possible sense. They are essential human roles in the responsible design and use of AI systems.
PO: Data gives AI its patterns. Human imagination gives it possibilities. Human judgment gives those possibilities direction.
8. Five Business Applications of Data-Borne Thinking
Application 1: Customer Experience
A bank collects customer support interactions.
Traditional analytics identifies common problems, response times, and satisfaction levels.
A generative AI system could help explore alternative customer journeys and communication approaches using appropriately protected information.
Provocation:
PO: Customer complaints are product design proposals.
Possible discovery:
Recurring confusion may reveal opportunities to redesign the service rather than merely improve complaint handling.
Application 2: Manufacturing
A manufacturing organization collects machine logs, maintenance records, and production information.
Predictive models may identify potential failures.
Generative tools could help engineers brainstorm maintenance procedures, alternative workflows, and training scenarios.
Provocation:
PO: Every machine failure is a design instructor.
Possible discovery:
Maintenance history can become an input to redesigning operational processes.
Engineering validation remains essential before any change is implemented.
Application 3: Education
Educational institutions maintain syllabi, learning materials, assessment results, and student feedback.
AI can help educators generate differentiated practice exercises, explanations, and alternative teaching approaches.
Provocation:
PO: The curriculum is not a document. It is a possibility generator.
Possible discovery:
Learning resources can become adaptable experiences rather than fixed sequences of content.
Application 4: Software Engineering
Software organizations possess code repositories, requirements, architecture documents, issue histories, and operational records.
Generative AI can help developers propose tests, explain unfamiliar code, draft implementations, and explore alternative designs.
Provocation:
PO: The codebase is a teacher of future software.
Possible discovery:
Existing software assets can support learning, design exploration, and developer productivity.
Generated code must still be reviewed, tested, secured, and maintained.
Application 5: Strategy and Innovation
Leadership teams frequently work with market research, performance reports, competitor information, and customer insights.
GenAI can help articulate competing scenarios, challenge assumptions, and generate strategic options.
Provocation:
PO: A strategy report is a laboratory of alternatives.
Possible discovery:
Instead of asking AI to summarize a report, leaders can ask it to identify assumptions, construct counterarguments, and explore unconventional strategic choices.
9. A Practical Framework: From Data to Accomplishment
How can organizations operationalize the data-borne perspective?
Consider the following six-stage framework, combining lateral thinking with practical AI use.
Stage 1: FOCUS — Identify the Opportunity
Choose a meaningful business problem. Define the desired outcome and identify which data is relevant and authorized for use.
Stage 2: PROVOKE — Challenge the Assumption
State an unconventional proposition.
Example: PO: Customer feedback should design the next product.
Stage 3: MOVE — Explore the Implications
Do not immediately accept or reject the provocation. Extract principles, identify differences, and explore where the idea could lead.
Stage 4: GENERATE — Use AI to Explore Alternatives
Use GenAI to develop multiple interpretations, scenarios, prototypes, and solution concepts.
Ask for genuinely different approaches, not simply variations in wording.
Stage 5: JUDGE — Evaluate the Alternatives
Assess feasibility, value, evidence, privacy, safety, cost, and organizational fit.
Distinguish plausible suggestions from unsupported claims.
Stage 6: ACCOMPLISH — Run a Small Experiment
Choose one promising possibility. Define success criteria, test it with appropriate human oversight, and learn from the results.
FOCUS → PROVOKE → MOVE → GENERATE → JUDGE → ACCOMPLISH
10. A Reusable AI Prompt: Discover Possibilities in Data
The following prompt can help managers, students, innovators, and AI practitioners move beyond routine data analysis.
Prompt: The Data-Borne Possibility Explorer
You are an innovation facilitator combining lateral thinking with Generative AI.
I will provide a business problem and a summary of relevant data or observations.
Your task is not merely to summarize or predict. Help me discover new possibilities.
- Identify the conventional assumptions behind the problem.
- Generate three provocative PO statements that challenge those assumptions.
- For each provocation, demonstrate lateral thinking movement.
- Develop at least five distinct possibilities.
- Explain which observations support each possibility and which claims remain speculative.
- Identify risks, limitations, and evidence needed.
- Recommend one small, low-risk experiment.
- Define measurable criteria for success.
Do not invent facts about the data. Distinguish evidence, interpretation, hypothesis, and recommendation.
Problem: [Insert problem]
Available data or observations: [Insert a non-sensitive summary]
Desired outcome: [Insert outcome]
11. The Risks of Data-Borne Thinking
The metaphor is powerful, but it must not encourage technological determinism.
Generative AI does not automatically transform data into trustworthy knowledge or useful innovation.
Several risks deserve particular attention.
Inherited Bias
Training data may contain historical biases, imbalances, and omissions. Generated outputs may reproduce or amplify them.
False Confidence
A fluent answer may be incorrect, unsupported, or misleading.
Privacy and Confidentiality
Organizations must ensure that sensitive information is used only through appropriately authorized and protected systems.
Intellectual Property
Training data, retrieved content, and generated outputs can raise copyright, licensing, and ownership questions.
Possibility Without Feasibility
An imaginative idea may be impractical, unsafe, economically unattractive, or inconsistent with real-world constraints.
Governance and Accountability
AI-generated recommendations require clear ownership, oversight, evaluation, and appropriate controls.
The central discipline is therefore:
Generate possibilities freely. Validate consequential decisions rigorously.
12. What This Means for Executive Leadership
The data-borne perspective suggests five questions for executive teams.
- Are we treating data only as a reporting asset? Could it also support product, service, and business-model exploration?
- Are our AI investments focused only on productivity? Where might AI help discover new sources of value?
- Are employees trained to question assumptions? Or are they mainly trained to request faster answers?
- Can we distinguish generated ideas from validated opportunities? Do we have a disciplined experimentation process?
- Who owns the consequences? Are accountability, governance, and human oversight built into the workflow?
The most important leadership shift may be from asking how many AI outputs can be generated to asking how many meaningful improvements can be demonstrated.
13. A Student and Practitioner Innovation Challenge
This idea can also become a practical learning exercise for students, freshers, and AI engineering teams.
Challenge: Turn a Dataset into Five Possibilities
Select a small public or synthetic dataset.
Examples include:
- Public transport delays
- Library book circulation
- Energy consumption
- Student learning preferences
- Customer feedback from a fictional business
Begin with a conventional analytical question.
For example: What causes delays in public transport?
Now introduce a provocation:
PO: A bus stop is a service, not a location.
Use GenAI to explore possibilities such as adaptive passenger information, alternative waiting experiences, or new service coordination models.
Then select one idea and prepare:
- A clear problem statement
- The original assumption
- The PO provocation
- The movement from provocation to idea
- A prototype or storyboard
- A simple evaluation plan
The objective is not to demonstrate that AI can produce five answers.
It is to demonstrate that the learner can transform information into a testable idea.
14. A Small Experiment for Every Reader
You do not need a large AI transformation programme to explore this proposition.
Choose one familiar report or dataset that you are permitted to use.
First, ask AI:
What are the three most important insights?
Next, ask:
What assumptions are we making about how this information should be used?
Finally, ask:
Generate three unconventional possibilities that could create value from this information. For each, propose a small experiment and explain what evidence would be needed.
Compare the outputs.
Did the third question produce a genuinely different direction?
Did any possibility deserve further investigation?
If so, you have experienced the practical difference between using AI for information and using AI for possibility exploration.
15. The Deeper Question: Is Intelligence Becoming Abundant?
For much of modern organizational history, analytical and creative expertise have been relatively scarce resources.
Generative AI makes certain cognitive tasks more accessible and less expensive.
It can draft, explain, summarize, translate, generate code, and explore alternatives at scale.
But greater availability of AI-generated output does not make sound judgment, trust, or successful execution automatic.
This leads to another provocation:
PO: When intelligence becomes abundant, the scarce resource is the ability to choose what matters.
The future of human work may depend less on producing every answer personally and more on framing meaningful questions, discovering valuable possibilities, and taking responsibility for results.
16. Conclusion: From Data to Possibilities to Value
The expression GenAI Is Data-Borne is not intended as a replacement for technical descriptions of generative AI.
It is a deliberate invitation to think differently.
Data-driven thinking asks us to learn from evidence.
Data-trained systems demonstrate how models can learn patterns from examples.
Data-borne thinking invites us to explore what new possibilities those learned patterns, combined with human purpose and context, might help generate.
The real opportunity is not the production of endless content or ideas.
It is the disciplined movement from information to imagination, from imagination to experimentation, and from experimentation to accomplishment.
Organizations will still need trustworthy data, capable systems, responsible governance, and sound economic judgment.
But they may also need something that conventional analytics programmes have not always emphasized:
A deliberate capability to generate and explore unconventional possibilities.
The Final PO
GenAI is data-borne.
Innovation is possibility-borne.
Value is accomplishment-borne.
Data provides patterns.
AI helps generate alternatives.
Human imagination challenges assumptions.
Human judgment selects directions.
Human action creates value.
17. Your Turn: A PO Thinking Invitation
Complete this statement:
PO: In the age of GenAI, data is no longer ________. It is ________.
Do not search immediately for the correct answer.
Generate a surprising interpretation. Explore where it leads. Identify one useful possibility.
That is the spirit of Lateral AI.
Further Reading and Conceptual Foundations
The following sources provide background for the established concepts discussed here. The expression "data-borne" and the proposed framework are exploratory interpretations, not terminology attributed to these sources.
- Edward de Bono, Serious Creativity (1992) — provocation, movement, and lateral thinking.
- Attention Is All You Need (2017) — the Transformer architecture. Read paper
- NIST, AI Risk Management Framework (AI RMF 1.0) — responsible AI risk management. Explore framework
- Stanford HAI, AI Index Report — research and evidence on AI developments. Explore reports
