Introducing the Management Consultant AI Operating System — Skills That Refuse What Would Hurt Your Client Trust

Update, September 2026: This pack is now the Management Consultant AI Operating System, at the same link and the same $19 price. It still includes every skill described below, and adds a command center and five jobs that run start to finish: win the work, kick off an engagement, weekly status and meeting follow-ups, the steerco pack or executive deliverable, and close and extend. If you connect Gmail, Google Calendar, and Google Drive, the jobs put drafts there for your approval; nothing sends without your click. Setup takes about two minutes, and a "Try it on a sample" run shows a full proposal before you paste anything real. The plugin now has 66 skills: 59 individual skills and 7 for the jobs and command center. It needs the Claude desktop app on a paid Claude plan, and every prompt also comes as a click-to-copy version for ChatGPT or any AI. The skill table below is updated to match. The rest of this post is the original May 2026 launch announcement.

There are a lot of "AI for consultants" prompt packs floating around. Most of them have the same problem: they're written by someone who hasn't sat through a real steerco at 11pm trying to defend a recommendation the CRO doesn't want to hear.

I just shipped a pack that fixes that. Management Consultant AI Prompts — 50 production-grade skills for working consultants. $19 one-time, lifetime updates, works in any AI.

You can grab it here: Management Consultant AI Operating System.

This post is about what's actually different.

The thing generic strategy GPTs get wrong

If you ask ChatGPT or Claude.ai web to "give me a Porter's Five Forces analysis," you'll get something that looks correct from 20 feet away. You won't get something you'd actually paste into a board deck.

The failure modes are predictable:

• Fabricated citations: a confident-sounding "per McKinsey Quarterly, March 2023" article that doesn't exist
• Mis-attributed frameworks: Porter's Five Forces dated 1985 (it's HBR 1979) or Minto's Pyramid Principle dated 2002 (it's 1987, expanded 1996)
• Invented case studies: "We helped a Fortune 500 retailer reduce churn 23%" — no client, no source, no provenance
• Misapplied frameworks: "Blue Ocean Strategy" name-dropped to mean "find a niche" without using the actual ERRC grid + value innovation principles
• Generic recommendations: "improve communication" as a deliverable

Any one of these in a board deck damages the consultant's credibility. Repeatedly, fatally.

What this pack does differently

Every skill is reviewed against a profession profile that documents the real source of every framework, regulator, and authoritative publication. The reviewer doesn't pass a skill until citations match the canon.

When the reviewer ran against the first 5 drafted skills, it caught:

• A Minto edition date that was off by one publication generation
• An Ansoff book title typo (one character — would've embarrassed anyone shipping it to a strategy client)
• An AARRR framework attribution that was anachronistic (the named org didn't exist yet in the cited year)

Zero of the 50 final skills shipped with fabricated sources.

The refuse-by-design feature

Generic prompt packs say yes to everything. This pack has skill-specific refuse conditions:

• m-and-a-target-screen refuses and routes to counsel if you paste material non-public information. The skill explicitly stops if your input includes a known-confidential deal.
• linkedin-post-from-engagement has a multi-layer NDA gate. It refuses to use the client name, deal size, attributable metrics, or any identifying detail. The skill walks you through an NDA-safety final check before posting.
• weekly-status-update refuses to soften a Red status into Amber to avoid escalation — surfaces the disagreement instead of complying.
• timesheet-narrative refuses to round up hours or pad billable narratives.
• engagement-letter-section refuses to draft indemnification, liability cap, or IP-assignment clauses — those route to counsel. Carries a "NOT LEGAL ADVICE" header in every output.

When a skill detects misuse, it surfaces the issue rather than quietly producing output that would damage your reputation.

What you get

Updated September 2026 to match the current plugin.

Five jobs and a command center (7 skills): win the work, kick off an engagement, send weekly status & meeting follow-ups, build the steerco pack or executive deliverable, and close & extend, plus the command center and a connect-your-tools skill.

59 individual skills, each of which still runs on its own:

Group	Count	Example skills
Setup & navigation	3	management-consultant-setup-wizard, management-consultant-skill-catalog, management-consultant-help
Always-on guards	2	confidentiality-guard, practice-boundary-guard
Problem structuring	4	mece-issue-tree, hypothesis-tree-draft, problem-statement-canvas
Diagnostics & research	8	industry-snapshot-brief, benchmarking-brief, m-and-a-target-screen
Strategy frameworks	6	porter-five-forces, bcg-growth-share-matrix, mckinsey-7s-diagnostic
Recommendation & synthesis	5	exec-summary-pyramid, one-pager-recommendation, board-deck-outline
Stakeholder management	4	stakeholder-map, steering-committee-prep, difficult-message-rewrite
Engagement cadence	5	weekly-status-update, change-request-memo, engagement-postmortem
Discovery & workshops	5	discovery-interview-guide, discovery-interview-summary, workshop-readout
Planning	3	roadmap-90-180-365, engagement-workplan-builder, risk-register-draft
Business development	5	proposal-cover-letter, engagement-letter-section, conflict-of-interest-check
Practice operations	5	timesheet-narrative, weekly-resource-plan, pipeline-review-prep
Personal brand & networking	4	linkedin-post-from-engagement, thought-leadership-outline, intro-email-warm-referral

A setup wizard that captures your practice profile (firm, jurisdiction, voice, frameworks you cite most often, words you never want in output). Every other skill reads it. Today it starts with a two-minute version: say "set me up" and paste your website, email signature, or a proposal you've sent.

In the download today:

• The plugin, for the Claude desktop app on a paid Claude plan
• A documentation pack: Start Here, a Jobs Guide, a one-page cheat sheet, and the skill catalog
• Every prompt as click-to-copy HTML, to paste into ChatGPT, Gemini, Copilot Chat, or any other AI

Who this is for

• Independent consultants billing $150-$500/hr running solo or 2-5 person practices
• Boutique-firm partners and principals
• Senior managers at Big 3 (McKinsey, BCG, Bain), Big 4, or Tier-2 firms (Strategy&, A.T. Kearney, L.E.K., Oliver Wyman)
• Post-MBA strategy consultants 1-5 years in
• Ex-consultants in corporate strategy / chief-of-staff roles

Not for: junior analysts without client interaction yet, or specialty consulting (M&A advisory, restructuring) — those domains have their own conventions.

Try before you buy

Two free options:

• The 90-Minute AI Audit — one prompt that scans your week and finds the 5 highest-leverage AI plays
• 7-in-1 Profession Prompt Sampler — one production prompt per profession; the consulting prompt is a representative example of what's in the full pack

If those click, the full pack is $19 with a 30-day refund. Email me if anything doesn't earn its keep.

— Alex Lowe, The AI Career Lab