Learning by Interrogation: An AI-Infused Pedagogy for Self-Directed Business Education

1. Introduction

As the world of business becomes faster, more complex, and less predictable, traditional education struggles to keep up. While most programmes deliver fixed content in a linear fashion, the real-world is neither fixed nor linear — it demands adaptability, systems thinking, and the ability to question assumptions — not just memorize them.

In this context, AI offers something powerful: not automation of teaching, but augmentation of thinking and learning.

This article introduces a pedagogy rooted in adductive reasoning, Socratic inquiry, and strategic foresight, all facilitated by AI. It reframes the role of learners: not as passive recipients of content, but as active model-builders, questioners, and decision-makers.

2. From Passive Absorption to Active Model-Making

Most AI-enabled learning today mimics traditional methods: give learners definitions, rules, or examples, and expect them to draw conclusions (induction) or apply known principles to cases (deduction).

These approaches assume knowledge is transferred in pieces, moving from simple to complex. But what if we reversed that?

Our model begins not with fragments, but with the whole — a draft business model, a messy scenario, or a flawed financial strategy. Learners don’t start by being told “how it works.”

Instead, they’re invited to make sense of the whole, then pull it apart, question it, and revise it.

This is the essence of adductive reasoning: starting with a possible explanation and testing its fit — starting with a well-formed question or problem statement crafted by the learner, which then generates or draws out a model through inquiry.

By combining adduction with AI-facilitated Socratic inquiry, we move learners through Bloom’s taxonomy in a nonlinear but accelerated fashion.

1. They analyze early (spotting gaps in flawed models),

2. Evaluate constantly (weighing scenarios and assumptions), and,

3. Explore iteratively (designing and refining future-fit models).

AI scaffolds this process by acting as:

• A model interrogator, asking "what if" and "why" at every turn,

• A feedback loop, revealing logical flaws or prompting deeper reflection,

• And a simulated peer or coach, offering diverse perspectives (investor, operator, skeptic).

The result is a pedagogy that doesn’t just support learning — it accelerates cognition, personalizes inquiry, and promotes future-ready strategic thinking.

Learners don’t just know the content. They think in systems, reason through ambiguity, and build with purpose.

3. The AI-Infused Pedagogical Approach

This approach is built on four key components, delivered through a learning loop that prioritizes reasoning over recall.

A. Adductive Reasoning as a Learning Entry Point

Rather than teaching a topic from the ground up, learners are shown or guided to create a full — even flawed — business model. This “whole first” approach (adduction) gives students a meaningful system to explore, stress-test, and rework.

Adduction is a reasoning process that begins with a well-crafted, learner-generated question or problem statement — a plausible challenge — which the learner then explores by iteratively building and refining a model.

This differs from:

• Induction: Building a rule from multiple examples

• Deduction: Applying a rule to a case

• Adduction (your model): Starting with a provocative or plausible problem → using AI and reasoning to construct and test a model to explain or address it.

This version of adduction mirrors how designers, entrepreneurs, and analysts actually think:

1. They start with a sharp question: “Why is this business failing despite revenue growth?”

2. Then they build a model: cash flow loops, team dynamics, pricing feedbacks, etc.

3. They refine that model with feedback (AI, peers, data).

AI becomes not just a tutor, but a collaborative explorer, surfacing weak logic, blind spots, and assumptions. This leads to...

B. Socratic Questioning via AI

AI plays the role of a relentless, curious coach, inviting, or asking, open-ended, probing questions like:

• “What if your main competitor adopts your strategy?”

• “Which assumption in your cost structure is most vulnerable?”

This style invites students to think, defend, doubt, and revise — not just consume.

C. Iterative Refinement with Feedback

Learners edit and justify their models through an AI feedback loop.

The AI identifies contradictions, prompts alternatives, and helps the learner revise — again and again — until the model begins to hold up under pressure.

D. Nurturing Strategic Foresight and Scenario Testing

Learners stress-test their models against potential future shocks:

• “What happens if interest rates rise 2%?”

• “What breaks first if your supplier vanishes overnight?”

They then create an MVFM — Minimum Viable Future Model — not just optimized for the present, but resilient in the future.

4. Why This Approach Differs from Conventional Classrooms

This pedagogy breaks from the content-first, quiz-later models of most business programmes. Instead:

• Students bring their own questions, creating intrinsic motivation and real ownership of their learning path.

• Multiple entry points (learners can choose what questions they want to start with) into each topic allow for diversity of thought and background — someone might start with pricing, another with operations.

• It creates a community of inquiry without needing to teach the theory of it. Learners naturally challenge one another, AI, and themselves.

• Each learner receives personalized guidance, since AI tailors prompts to their unique thinking.

• Students leave not just with knowledge — but with mental agility, strategic foresight, and experience interrogating real business complexity.

Most classrooms, although touted as "learner-centred", are still very facilitator oriented except for Inquiry based learning. As a result, a level of co-dependence between facilitator and the class still persists. This approach increases the proportional role of the facilitator as facilitator or mentor with the possibility of higher levels of self-directed learning in future.

From early experiments, these key benefits emerged:

4. Why This Approach Works Because of AI (Not Just with It)

Traditional learning methods weren’t designed with AI in mind — they rely on preset content, linear progression, and one-size-fits-all explanations. But your pedagogy — rooted in adductive inquiry — is tailor-made for AI’s unique strengths:

1. AI Is Exceptionally Good at Responding to Questions, Not Lecturing

Unlike a textbook or video, an AI doesn't need a prewritten sequence. It responds dynamically to the learner's question or model — probing, elaborating, or challenging based on the learner’s direction.

This means learners can begin anywhere: their own question is the curriculum.

2. AI Doesn’t Have “Correct” Endpoints — It Supports Exploration

AI thrives in gray zones — helping learners explore trade-offs, assumptions, “what if” scenarios, and multiple valid paths. It supports reasoning over recall, and inquiry over instruction.

This makes it ideal for scaffolding thinking up Bloom’s taxonomy, rather than drilling facts at the bottom.

3. AI Can Play Roles Instantly — Coach, Critic, Investor, Customer

Using prompts like “Act as an investor. What concerns would you raise?”, learners get immediate, persona-based feedback — something human teachers or peers can’t scale in real time.

This simulates a community of inquiry — from just one learner and one interface.

4. AI Responds in Natural Language — Matching Learner Voice

Because AI understands questions in the learner’s own words, it reduces the intimidation barrier. Learners don’t need to “know the terminology” first — they can think out loud, and the AI helps shape that into structure.

This makes complex systems thinking more accessible — even to novices.

5. AI Makes Reflection Instant and Iterative

At any point, learners can ask:

• “What assumptions am I making?”

• “What are the unintended consequences?”

• “What’s missing in my model?”

This makes metacognition a natural part of the loop, not a separate module.

As a corollary, learners can interrogate the products of their inquiries according to their own needs. We have also seen instances where learners asked AI to tune the language to their appropriate level such as IELTS Level 3 or 4. This is despite their level of language proficiency as a simplified text often clarifies the main concepts, especially the jargon.

In other words, this pedagogy — where learners start with their own question, then build and refine a model through AI-facilitated Socratic inquiry — isn’t just a novel teaching method. It’s a pedagogy native to AI.

It’s fast, flexible, personalized, and cognitively rich — not because it uses AI to deliver content, but because it uses AI to think.

6. Prerequisites for Success

From early experiments, several key benefits emerged:

The emerging question is then presented as:

How much does this model thrive on prior experience?

However, we are clear that the experience doesn’t have to be formal:

A. Activate Prior Knowledge

Even informal or non-formal exposure to business ideas (budgeting, selling, project planning) provides a base. Encouraging students to reflect on what they already know makes new learning “stick.”

B. Minimum Viable Entry

You don't need MBA-level learners. The lowest effective floor:

• Secondary school level (~age 15+)

• Ability to type, read, and reflect

• Willingness to ask “why” and revise ideas

We encouraged learners to climb the ladder of What, Where, When, Who, Why and How.

C. Teach AI Literacy Early

Before serious inquiry begins, spend 30–60 minutes introducing:

• Good prompt structure

• Follow-up questioning (especially the Socratic Approach)

• Reflection on AI’s assumptions and reasoning

At its core, Socratic learning is not about giving answers — it's about asking the right questions to provoke deeper thinking. It includes:

• Challenging assumptions

• Encouraging self-explanation

• Exploring contradictions

• Pushing the learner to clarify, justify, and revise

AI becomes Socratic when it’s used to:

• Ask instead of tell: "Why do you think this revenue model is sustainable?"

• Interrogate logic: "What assumption are you making about customer behaviour?"

• Surface risk: "What breaks first if your cost doubles overnight?"

• Play critical roles: "I'm your COO. What are the operational risks here?"

In this pedagogy, learners flip the script:

• They prompt the AI to question them (“Act like a skeptical investor—what would you challenge?”)

• They reflect on the AI’s reasoning (“What assumptions are you making in this explanation?”)

• They use it to generate multiple perspectives on one idea

This transforms AI from an answer engine into a thinking partner — the essence of Socratic learning.

7. Use Prompts That Build Models, Not Lists

For example, do not ask, “What is product-market fit?” but rather use, “Help me build a model that shows what happens if my product doesn't meet market demand.”

Why: This encourages systems thinking, not just fact recall.

End with Synthesis

Prompt:

“Summarize the contradictions we uncovered and what I should rethink.” “Based on our exchange, how could I redesign my pricing model for more stability?”

Why: Learners consolidate insights and move from analysis to creation (top of Bloom’s taxonomy).

Summary Flow

We present here a Mnemonic to guide the prompt sequence and flow: "CRAFT IT"

Each letter maps to an intentional phase of the AI–learner interaction loop:

Readiness AreaWhat’s NeededHow to Build ItPrior KnowledgeSome personal or observed experienceUse analogies (e.g., home budgeting as cash flow)CuriosityWillingness to explore and reviseModel vulnerability and “safe uncertainty”AI FluencyBasic interaction with prompts and repliesMicro-lesson on prompt writing and reflectionSystems ThinkingUnderstands cause-effect dynamicsUse simple loops (e.g., late payments → cash strain)

8. Implementation & Assessment

Delivery Options

• Micro-sprint (2 hours): One concept + AI-guided inquiry + peer debrief.

• Modular Programme (6 weeks): Build from basic loops to capstone simulation.

Assessment Options

9. Conclusion

AI is often seen as a threat to teaching — but when used right, it becomes a Socratic companion. By integrating adduction, interrogation, foresight, and personalized AI scaffolding, we can create business learners who don't just know what to think, but know how to think critically, systemically, and for the future.

This way, the learner’s own question triggers their upward movement through Bloom’s levels — with AI acting as a cognitive scaffold, not an answer machine — as set out in the table below.

How Adduction Accelerates Bloom’s Taxonomy

Here’s an updated pyramid that reflects this flow:

This isn’t just a new course design — it’s a new stance on education: one that sees thinking as the skill, not content as the goal.

“AI is like a relentless two-year-old: it keeps asking ‘Why?’ — and that’s exactly what mature learners need.”