Applying Consulting Frameworks with AI

Frameworks are a consultant's grammar — Porter's Five Forces, 7S, MECE issue trees, 2x2 matrices, value chains, BCG growth-share. AI is excellent at applying these frameworks to a specific situation, faster and often more rigorously than a tired consultant at midnight.

This lesson shows you how to use AI as a thinking partner — not just a content generator — for the analytical core of an engagement.

What You'll Learn

• How to use AI to apply standard frameworks to a specific client
• Building MECE issue trees and hypothesis maps with AI
• Using AI as a structured "second opinion" on your analysis
• The framework prompts that work best in practice

The Framework-as-Prompt Pattern

Most consulting frameworks have a known structure. AI handles structure beautifully. The pattern is always the same: tell the AI which framework, give it the situation, ask it to apply the framework rigorously, and demand specifics.

A good framework prompt has four parts: framework, context, depth requirement, output format.

Porter's Five Forces

Apply Porter's Five Forces to the European mid-market HR-tech sector, from the perspective of a regional retail bank that is considering acquiring a player in this space. For each force: (1) rate intensity High/Medium/Low, (2) give 3 specific evidence points (with examples of real companies if relevant), (3) state the strategic implication for our client. Use a markdown table.

What makes this prompt strong is the perspective ("from the perspective of a regional retail bank") and the evidence requirement. Without those, AI gives a textbook answer that could apply to any sector.

MECE Issue Trees

Issue trees are the backbone of structured problem solving. AI is exceptional at the first draft.

Build a MECE issue tree for this question: "How can ClientCo improve operating margin from 12% to 18% within 24 months?" ClientCo is a $2.4B North American specialty retailer with 480 stores and a small but growing e-commerce channel. Use 3 main branches (e.g., increase revenue, decrease COGS, decrease SG&A) and 3-4 sub-branches per branch. For each leaf, state the specific lever and one example of the analysis we would run to test it. Confirm at the end that the tree is mutually exclusive and collectively exhaustive — if not, fix it.

The "confirm at the end" line forces the AI to self-audit. Often the first version has overlapping branches; the audit catches them.

2x2 Matrices

I am building a 2x2 to prioritize 14 strategic initiatives for a CPG client. Suggest 5 candidate axis pairs that would create insight. For each pair, describe what each quadrant means and which kind of decision the matrix supports.

Then once you pick axes:

Plot these 14 initiatives on the matrix with axes "Strategic value (low–high)" and "Ease of implementation (low–high)". For each initiative, give a brief justification for placement. Be willing to place initiatives in unfavorable quadrants — do not bunch everything in the top-right.

The "do not bunch everything in the top-right" instruction matters — without it, AI tends to be diplomatic and useless.

The Hypothesis-Driven Approach

Top-tier strategy firms work hypothesis-first: they form a sharp answer early, then test it. AI makes this faster.

Based on the discovery interviews and market research below, propose 3 distinct hypotheses for why ClientCo's customer churn has risen from 8% to 14% over the last 18 months. For each hypothesis: (1) the one-sentence claim, (2) the 3 strongest pieces of supporting evidence, (3) the 2 strongest pieces of disconfirming evidence, (4) the analysis we would run to test it conclusively. Output as markdown with each hypothesis as its own section.

Discovery: [paste] Research: [paste]

Then sharpen with a counter-prompt:

Now act as a contrarian senior partner. For each hypothesis, identify the single most likely reason it is wrong. Then propose a fourth hypothesis that none of the above three considered.

The fourth hypothesis is often the most interesting, because it forces AI past the obvious.

Using AI as a Structured Second Opinion

Beyond drafting, the highest-leverage use of AI in analysis is as a critic. After you build your own framework or analysis, run it past AI:

Below is my Five Forces analysis for ClientCo. Critique it from three angles: (1) where am I being too generic — what specifics am I missing? (2) where am I conflating two distinct issues? (3) what important force or dynamic am I underweighting? Be direct, not polite.

[paste your analysis]

This is the prompt that most consultants underuse. AI as a critic is faster and often sharper than a peer review, and it never tells you the analysis is "great work."

Framework Library: Prompts Worth Saving

Build a personal prompt library. The frameworks below are the ones consultants reach for most often — keep a working prompt for each, ready to paste and adapt.

• SWOT — but always with "for what decision?" specified
• Porter's Five Forces — always with a perspective
• 7S model — useful for organization assessments
• Value chain — useful for cost or margin analysis
• BCG growth-share matrix — useful for portfolio decisions
• Ansoff matrix — useful for growth strategy
• Job-to-be-Done — useful for product or service redesign
• Customer journey map — useful for CX and operations work
• PESTEL — useful for macro/regulatory framing
• MECE issue tree — useful for any "how do we..." question

For each, store a prompt template in a notes app. After 30 engagements, you have an asset that compounds.

Common Pitfalls

• Do not let AI pick the framework for you. A framework is a perspective. The choice belongs to the consultant.
• Do not paste a generic AI framework into a deck. Use AI's output as raw material; rewrite in your firm's voice.
• Do not skip the contrarian round. First-draft AI analysis tends to be balanced and forgettable. Force a contrarian view.

Key Takeaways

• The framework-as-prompt pattern: name the framework, give context with a perspective, demand specifics, and define the output format.
• Issue trees, 2x2s, and Five Forces work especially well — but always include self-audit and contrarian prompts.
• Use AI as a structured critic on your own analysis — this is one of its highest-leverage uses.
• Build a personal prompt library indexed by framework. It compounds over your career.