By the end of this walkthrough you will have a repeatable engagement rhythm: problem frame → source library → structured analysis → client artifact — the same pipeline senior associates run, with less re-collection between projects.
Consulting market research breaks down in predictable places: cold-start (new sector, 48 hours to workshop), mid-deck source requests (partner asks "where's the citation?" at slide 14), and Monday brief pressure (seventeen headlines, two matter). AI does not fix these by summarizing more — it fixes them when research lives in a persistent, structured library you query and extend.
What you'll need
- A client or internal mandate with a defined decision (enter market, pricing move, M&A target screen)
- 5–10 seed sources: industry reports, competitor sites, expert transcripts, prior engagement notes
- Mindar (or equivalent) with knowledge libraries and analysis skills
- 2–4 hours for initial setup; 30–45 minutes daily for maintenance
Step 1: Boundary the problem (Frame)
Before opening a browser tab, write four lines:
- Decision: What must the client decide?
- Scope: Geography, segment, time horizon
- Success criteria: What would a "good enough" answer include?
- Known unknowns: What you explicitly will not answer this week
In Mindar, open a new library named after the mandate. Add these four lines as the first note — every later clip should connect back to them.
Why this matters: Without a frame, AI produces fluent market overviews that do not sharpen the client's decision.
Step 2: Build the source library (Collect)
Add sources in layers:
| Layer | Examples | Purpose |
|---|---|---|
| Market structure | Analyst reports, trade press, regulatory filings | Size, segments, growth |
| Competitive set | Pricing pages, product docs, job posts | Positioning, capability |
| Customer voice | Reviews, forum threads, interview notes | Jobs-to-be-done, objections |
| Internal | Prior decks, workshop notes | Continuity, client context |
Tag every item: segment, competitor, claim-type (fact vs. opinion vs. forecast).
Time to first result: After 5–10 sources, ask: "Who are the top 5 players by segment and what is each one's wedge?" Verify citations against your library.
Step 3: Run structure skills before narrative (Analyze)
Use analysis patterns that match how partners think:
- Market map — who plays where, before the deck fills with bullets
- Driver tree — what actually moves the thesis (not every trend line)
- Hypothesis board — explicit bets you are testing with evidence
- Source compare — same claim across two sources; flag conflicts
Run these against your library, not the open web. The output is scaffolding for slides and memos, not a substitute for judgment.
Step 4: Filter variables and compare options (Decide)
List the 3–5 variables that swing the recommendation (price, switching cost, regulation, incumbency). For each:
- State the current best estimate
- Cite the strongest source in your library
- Note confidence (high / medium / low)
Build a simple options matrix: rows = strategic choices, columns = variables. Gaps become your research agenda for week two.
Step 5: Ship the client artifact (Deliver)
Pick one primary format based on the meeting:
- Workshop pre-read — frame, market map, 3 hypotheses, open questions
- Decision memo — recommendation, evidence, risks, what would change our mind
- Comp matrix — feature/capability grid with source footnotes
Draft in Content Factory from the library; verify every claim before the client sees it. Partners trust associates who show passages, not paraphrases.
What you built
You now have:
- A mandate-scoped library reusable across phases of the engagement
- Structured artifacts (maps, trees, matrices) that update when sources change
- A delivery template your team recognizes
Next steps: Subscribe to a Plaza starter library for your sector, or clone this library as a template for the next mandate.