Using AI to Draft, Critique & Tighten a Deck
A weak deck does not fail because the fonts are ugly. It fails because slide 3 says “market is attractive,” slide 4 says “competition is intense,” and nobody can tell what decision the deck is driving.
AI can now produce ten slides in seconds. The real skill is making sure those slides have a sharp answer, clean logic, credible evidence and no consultant-sounding fog.
- Use AI as a thinking partner, not a slide factory. You own the judgment, storyline and final recommendation.
- A strong deck follows one argument: answer first, reasons second, evidence third.
- The best workflow is draft - critique - tighten: generate structure, attack the logic, then compress the message.
- Never ask AI to “make a deck.” Ask it for slide titles, storyline options, evidence gaps and objections.
- Every slide needs an action title: a sentence that states the insight, not a topic label.
- The biggest risk is polished nonsense: fluent slides with weak facts, vague causality or unsupported recommendations.
- For interviews, explain your AI use ethically: “I used AI to stress-test structure and wording, but verified the logic myself.”
Big Picture: AI Is the Junior Analyst, Not the Partner
Think of AI as a fast junior analyst who can produce options, spot inconsistencies and rewrite messy slides. But it does not know your client, cannot guarantee facts, and should never decide the recommendation for you.
Core Explanation: The Draft - Critique - Tighten Loop
Most students use AI too early: they paste rough analysis and ask, “Create a deck.” That produces a deck-shaped document, not a business argument. The better approach is to separate the work into three jobs.
1. Draft: Get the Skeleton Before the Slides
Start with the decision, audience and constraints. A consulting-style deck is built backwards from the answer: what does the client need to decide, what recommendation are you making, and what proof will make it credible?
If your problem itself is fuzzy, fix that first. AI becomes far more useful after you have clearly framed the business question; this is why defining the problem before solving it is the natural prerequisite to any AI-assisted deck.
2. Critique: Make AI Attack the Deck Like a Skeptical Partner
Once you have a draft storyline, do not ask AI to “improve” it. Ask it to find what is wrong. Give it a role: partner, CFO, category head, investor, operations leader or interviewer.
Useful critique prompts:
3. Tighten: Reduce Words Without Reducing Meaning
Tightening is not beautification. It is compression. A tightened deck removes duplicate points, vague words, unsupported adjectives and “analysis tourism” - slides that are interesting but not necessary for the decision.
Definitions: The Terms You Should Say Cleanly
- Deck: A sequence of slides that carries a business argument toward a decision.
- Action title: A slide headline that states the insight or implication, not just the topic.
- Storyline: The ordered logic connecting the recommendation, supporting reasons and evidence.
- Pyramid Principle: Barbara Minto’s method of presenting the answer first, then grouped supporting arguments (Barbara Minto).
- MECE: A grouping standard where points are mutually exclusive and collectively exhaustive.
Mini Case Study: McKinsey’s Lilli and the New Deck-Builder Mindset
McKinsey built Lilli as a generative AI platform to help consultants access firm knowledge and accelerate synthesis, which shows where AI fits in professional problem-solving.
Consulting work has always had a hidden bottleneck: not just doing analysis, but finding relevant prior knowledge, synthesizing it, turning it into a client-ready narrative and pressure-testing the recommendation. McKinsey describes Lilli as its generative AI platform that helps colleagues access and use the firm’s knowledge.
The strategic move is important: Lilli is not positioned as “press a button, get a final answer.” It supports retrieval, synthesis and drafting. The consultant still has to frame the problem, interpret trade-offs, validate facts and own the client recommendation.
The lesson for an MBA student is direct. Use AI the same way: first to search your own notes, then to propose storyline alternatives, then to critique slide logic, then to tighten language. The primary driver is faster synthesis. The supporting drivers are better retrieval of prior material, more critique cycles and cleaner first drafts. The human driver remains judgment.
If you were preparing a deck on Zomato’s food delivery and quick-commerce choices, AI could help convert annual-report notes and public commentary into storyline options. But the recommendation would still depend on Indian market realities: dense urban demand, delivery economics, competitive intensity, restaurant partner relationships and customer frequency. The so what: AI can arrange the argument, but you must understand the business model.
How AI Changes Using AI to Draft, Critique and Tighten a Deck
By 2026, AI changes deck-making in three practical ways.
1. From Slide Creation to Storyline Simulation
Earlier tools helped you design slides. Newer AI workflows help you simulate alternate arguments: “Should the deck lead with market attractiveness, profitability risk or execution feasibility?” This is powerful because consulting interviews reward structured judgment, not decoration.
2. From Grammar Checking to Executive Critique
AI can role-play different stakeholders. A CFO may attack payback assumptions; a sales head may challenge channel feasibility; an interviewer may ask why your recommendation beats the next-best alternative. This helps you find weak points before the real discussion.
3. From Blank-Page Work to Evidence-Aware Drafting
Tools can summarize transcripts, company filings, research notes and your own analysis. But this creates a new danger: AI may blend verified facts with plausible filler. Your job is to maintain a fact log and label every claim as verified, assumption or hypothesis.
Load your case notes, Excel outputs and company annual report into NotebookLM. Ask: “Create a 10-slide storyline with action titles, mark evidence gaps, and list five partner-style objections.” Then verify every claim before you use it.
If you want to practise the interview side of this workflow, pair deck creation with using AI as a mock interviewer so the same recommendation is tested verbally.
Interview Relevance
“Suppose you have completed the analysis for a market-entry case. How would you use AI to create and improve the final client deck?”
Use the phrase: “AI can accelerate synthesis, but accountability for judgment stays with the consultant.” That line signals maturity.
Common Mistake
The mistake: using AI to make the deck look polished before the argument is true. It costs candidates because interviewers quickly spot unsupported recommendations, vague slide titles and missing trade-offs. One-line fix: lock the recommendation and logic tree first, then use AI only to draft, critique and tighten.