Claude Prompts for Management Consultants: 18 Templates for Proposals, Memos, and Client Reports
It's 9pm on a Sunday. The steerco deck is due Tuesday morning. You have notes from four interviews, a half-built financial model that needs commentary, and the partner's voicemail from the airport saying "let's tighten the recommendation slide before I land."
Every consultant knows this hour. Until recently, the only options were caffeine and time. Now there's a third one — and if you're using ChatGPT for it, you're leaving a lot on the table.
Claude is the better tool for consulting work. Not because of marketing copy, but because of three real differences: a 200K-token context window that swallows whole annual reports and interview transcripts without breaking a sweat, native handling of the structured XML-style instructions that most consulting prompts depend on, and a markedly better instinct for the "so what?" filter that turns research into a recommendation. The output reads like a draft you'd actually edit, not one you'd rewrite.
Below are 18 Claude prompts built around the actual artifacts a consultant produces in a week: proposals, executive summaries, client status reports, interview synthesis, workshop guides, and the small-but-painful "translate this for the client" tasks. They're written for Claude specifically — most use XML tags and the structured-instruction pattern Claude handles better than other models.
Two setup notes before the prompts.
Setup notes
Confidentiality first. Most consultants are subject to NDAs that prohibit pasting client-confidential information into consumer AI tools. Claude for Work / Claude Enterprise has a zero-retention setting and doesn't train on your inputs — that's the version to use for client material. If you're on the free tier, redact client names, financial figures, and proprietary frameworks before pasting. The prompts below all use placeholder variables ([CLIENT], [INDUSTRY], etc.) for that reason.
The "consultant mode" preamble. Paste this once at the top of any Claude session and reuse the chat. It changes the output quality more than any individual prompt:
You are a senior consultant at a top-tier strategy firm
(McKinsey/Bain/BCG-caliber). Apply these defaults to
everything you produce in this session:
- Pyramid principle (Minto): governing thought first, then
supporting arguments, then evidence
- 80/20 instinct: lead with what matters, cut what doesn't
- "So what?" filter: every paragraph earns its place by
answering "and therefore?"
- MECE structure where lists are involved
- Plain English, not consultant jargon. No "synergies,"
"leverage," "unpack." Direct verbs.
- When the input is ambiguous, ask one sharp clarifying
question instead of guessing. When you do guess, flag it.
Be willing to push back if my framing is weak. Don't agree
just because I'm the one asking.
That last instruction — "be willing to push back" — is the difference between Claude as a fast typist and Claude as a junior consultant who actually improves your thinking.
Now the prompts.
Proposals & Engagement Letters
1. The proposal first-draft prompt
The hardest part of any proposal is starting. This gets you to a 70% draft.
Draft a consulting proposal for the following opportunity.
<context>
Client: [INDUSTRY, size, ownership type]
Their stated problem: [WHAT THEY SAID THEY WANT]
The real problem (your read): [WHAT YOU THINK IT ACTUALLY IS]
Decision-makers: [WHO READS THIS — CEO, board, ops VP?]
Competitors for the work: [IF KNOWN]
Our differentiated angle: [WHAT MAKES US THE RIGHT CHOICE]
Budget signal: [IF ANY]
Timeline: [WEEKS/MONTHS]
</context>
Structure the proposal as:
1. Executive summary (one page, governing thought up front)
2. Our understanding of the situation (show we've listened)
3. Proposed approach — phases, key activities, deliverables
4. Why us (case examples, team strengths — keep tight)
5. Investment and timeline
6. Next steps
Tone: Confident, direct, no inflated language. Read like a
partner wrote it, not a marketing team. About 8-10 pages of
content when formatted.
2. The "engagement letter from a proposal" prompt
Once the verbal yes is in, the engagement letter has to lock the scope without sounding like a contract from a law firm.
Convert the attached proposal into an engagement letter.
Keep:
- Scope (what we're doing)
- Out of scope (what we're not — be explicit)
- Deliverables and dates
- Fees, expenses, and payment terms
- Client responsibilities (data access, stakeholder time)
- Change-control process
- Confidentiality and IP
- Term and termination
Tone: Business-like, not legalistic. Keep clauses short and
in plain English. The client should be able to read this and
sign without forwarding to legal — though they probably will.
<proposal>
[PASTE]
</proposal>
3. The proposal-pricing sanity check
Before you send a fee number, run it through this.
I'm proposing a fee of [AMOUNT] for the engagement described
below. Stress-test this number.
<engagement>
Scope: [PASTE FROM PROPOSAL]
Team: [SENIORITY MIX AND WEEKS PER PERSON]
Duration: [WEEKS]
Expected deliverables: [LIST]
</engagement>
Tell me:
1. Implied blended rate. Is that defensible for the
seniority mix and the work?
2. Where the scope risk is — what's most likely to expand
mid-engagement?
3. What I should explicitly call out as out-of-scope to
protect the margin
4. Whether the fee is too low (signal of inexperience), too
high (signal of poor fit), or in the zone
5. One alternative pricing structure (fixed-fee with
milestone gates, T&M cap, value-based) and when I'd
choose it instead
4. The "lost the deal" debrief prompt
When a proposal loses, the temptation is to move on. Don't. This extracts the lesson.
We lost the proposal described below. The client's stated
reason was [WHAT THEY SAID]. My honest read is [WHAT YOU
THINK].
<proposal>
[PASTE OR SUMMARIZE]
</proposal>
Help me debrief. Specifically:
1. What signals in the original brief did we miss or
misinterpret?
2. Where does our written proposal probably underperform a
competitor's? Be specific. Don't be polite.
3. What's one thing about our approach that a buyer might
have read as a red flag (e.g., team too junior, scope
too rigid, fee structure mismatched to their procurement)?
4. What would I write differently if I had this proposal
back tomorrow?
Goal: a one-page memo I can share with the practice partner.
Executive Summaries & Strategy Memos
5. The one-page executive summary prompt
The piece every client actually reads. This nails the structure.
Write a one-page executive summary for the deliverable below.
Apply Minto's pyramid principle: lead with the governing
thought (the answer to "what should we do?"), then 3-4
supporting arguments, then evidence underneath each.
<deliverable>
[PASTE FULL FINDINGS — can be 30+ pages]
</deliverable>
<context>
Audience: [CEO / board / ops team]
Their main question: [THE ONE THING THEY ARE DECIDING]
Current state of the recommendation: [DRAFT / FINAL]
</context>
Constraints:
- One page max (about 350 words)
- The first sentence is the answer to the question. Not
background. Not "this report covers..."
- Three to four supporting arguments. Each is a complete
sentence, not a label.
- Evidence is specific (numbers, examples) not general.
- No "leverage," "synergy," "unpack," "moving forward."
Stop and ask if the audience or the main question is unclear.
6. The strategy memo from messy notes
You ran a 90-minute strategy session. You have whiteboard photos and a 14-page transcript. This converts both into a structured memo.
Below are notes from a strategy session with [CLIENT].
Convert them into a 3-page strategy memo.
<notes>
[PASTE TRANSCRIPT, BULLETS, OR DICTATED NOTES — anything goes]
</notes>
<context>
The decision in front of the client: [E.G., "WHETHER TO
ENTER THE EU MARKET IN 2026"]
The audience for this memo: [E.G., "CEO + 2 BOARD MEMBERS"]
What I want this memo to drive: [E.G., "A YES/NO DECISION
AT THE NEXT BOARD MEETING"]
</context>
Structure:
1. The question (one sentence)
2. Recommendation (one sentence, taking a side)
3. Why — three reasons, each with a paragraph
4. Risks and how we'd mitigate them
5. What we'd need to decide today vs. later
6. Open questions for the next conversation
Take a position. If the notes are split, name the split and
recommend the side you think is stronger, with reasoning.
Do not output "it depends."
7. The "two-pager" working document
When the client wants a working document, not a finished deck.
Draft a two-page working document for [CLIENT_TEAM] on
[TOPIC].
<inputs>
- Recent client interviews: [PASTE OR ATTACH]
- Internal hypothesis going in: [STATE IT]
- What we learned that contradicted it: [LIST]
- What we learned that confirmed it: [LIST]
</inputs>
Format the two-pager as:
- Page 1: What we know now (revised hypothesis, key
evidence, what changed)
- Page 2: What we'd test next (3 options, each with
effort/value/timeline)
Voice: Internal working tone. Not pitched. Not polished.
The goal is to provoke a useful conversation in tomorrow's
session.
8. The recommendation slide rescue
The deck has 47 slides. Slide 38 is the recommendation. It's not landing. This rewrites it.
Below is the current recommendation slide and the supporting
context. The client's reaction in the dry run was
[FEEDBACK — e.g., "they got lost halfway down" or "they
didn't see why we chose option B"].
<current_slide>
[PASTE TITLE, BULLETS, AND ANY CHART CAPTIONS]
</current_slide>
<context>
[PASTE THE 2-3 SLIDES BEFORE IT FOR CONTEXT]
</context>
Rewrite the slide so:
- The title is a complete sentence stating the
recommendation. Not a label like "Recommendation."
- The body has at most 3 supporting points.
- Each point starts with the implication, not the evidence.
- The "why not the other options" logic is visible in one
line, not buried.
Give me 2 versions: one tight (single-page), one expanded
(split across this slide and a follow-up slide if the
recommendation needs more air).
Client Status Reports
9. The weekly client update email
The Friday email that keeps the engagement on track.
Draft this week's client update email.
<this_week>
What we did: [LIST]
What we learned: [KEY INSIGHTS]
Decisions made: [LIST]
Decisions pending: [LIST WITH OWNER]
Risks emerging: [LIST]
</this_week>
<next_week>
Planned activities: [LIST]
What we need from the client: [LIST WITH NAMES AND DATES]
</next_week>
Format:
- Subject line that signals the headline (e.g., "Week 3 —
pricing test results in, decision needed by Tuesday")
- 3-4 short paragraphs, not bullet hellscape
- The asks for the client are one clear bulleted list at
the end, with names and dates
- Under 350 words
Tone: Direct, partner-of-record voice. Not breezy. Not
formal-formal.
10. The steerco prep memo
For the bi-weekly steering committee, a one-page prep doc for the partner.
Draft a steering committee pre-read for our
partner-in-charge.
<engagement_status>
Engagement: [NAME], Week [#] of [#]
Workstreams: [LIST WITH STATUS]
Findings to date: [TOP 3]
Decisions needed at this steerco: [LIST]
Open risks: [LIST WITH OWNER]
Client mood / political read: [NOTES]
</engagement_status>
Format as a one-page memo:
1. The 3 things that must come out of this meeting (in
priority order)
2. The 2 questions the partner should expect from the
client side
3. The 1 ask we want from the client
4. Anything political or fragile that the partner should
know but not say out loud
Last section is for the partner only — flag it as such.
11. The "explain the delay" client email
When something slips, the email has to acknowledge without grovelling.
Draft an email to the client explaining a slip.
<situation>
What slipped: [DELIVERABLE / MILESTONE]
By how much: [DAYS/WEEKS]
Why: [REAL REASON — could be data delay, key person out,
scope expansion, our resourcing]
Mitigation: [WHAT WE'RE DOING ABOUT IT]
New committed date: [DATE]
</situation>
Constraints:
- No "we apologize for any inconvenience"
- One sentence acknowledging the slip
- One sentence on the cause, in plain language
- The mitigation is the heart of the email
- Specific new date, not "shortly"
- Under 150 words
The relationship is [HEALTHY / TENSE / NEW]. Adjust tone
accordingly.
Interview & Research Synthesis
12. The interview synthesis prompt (across multiple transcripts)
Six interviews, 90 minutes each. The synthesis slide is due tomorrow.
Below are transcripts from [N] interviews conducted as part
of the [ENGAGEMENT] discovery phase.
<transcripts>
[PASTE — Claude's 200K context handles this; just dump them
with clear delimiters between each]
</transcripts>
<framing>
Hypothesis going in: [STATE IT]
Audience for the synthesis: [PARTNER / CLIENT / INTERNAL]
Decision the synthesis informs: [STATE IT]
</framing>
Output:
1. The 4-5 themes that emerged across interviews. For each:
- Theme statement (a sentence, not a label)
- How many interviewees raised it (count + roles, no names)
- One representative quote (paraphrased, not verbatim —
for anonymity)
- Whether this confirms, complicates, or contradicts the
hypothesis
2. The 2-3 outliers — things only one or two people raised
but that feel important
3. The questions we should now go answer that we didn't ask
in round 1
4. A draft governing thought for the synthesis page, if you
were going to write the recommendation today
Be honest where the evidence is thin.
13. The industry primer prompt
When you get staffed on a new sector at 5pm on a Friday.
I just got staffed on [INDUSTRY / SUB-SECTOR]. I have until
Monday to be fluent enough to interview a CEO. Build me a
crash primer.
Include:
1. Industry structure: who plays, value chain, where the
profits sit
2. The 5-7 things that differentiate winners from losers in
this space
3. The major transitions in flight (regulatory, tech,
consumer, capital flows)
4. The 10 acronyms / terms that an outsider mispronouncing
would immediately flag them as a tourist
5. The 3 most-quoted recent reports or data sources in this
industry — what they say and where they're contested
6. 5 sharp questions I should ask a CEO that signal I've
done my reading
7. 3 honest blind spots in your own knowledge — areas where
I should validate with a primary source
Be willing to say when something is changing fast and your
information might be dated.
14. The "pulled from the annual report" research dump
For the work where you do need to read the 10-K but not all 280 pages.
I've attached the 10-K for [COMPANY]. I'm preparing a
proposal/engagement involving [TOPIC — e.g., "their cost
position in their largest segment"].
Extract only what's relevant:
1. Revenue and margin trajectory by segment, last 3 years
2. Stated strategic priorities (with page references)
3. Disclosed risks that touch the topic above
4. Capital allocation patterns (M&A, capex, buybacks,
dividends)
5. Anything in MD&A or the footnotes that suggests stress
not yet reflected in the headline numbers
6. A list of 5-7 questions the 10-K raises that a
consultant would want to dig into
Cite page or section for every claim. If the document
doesn't support a claim, say so.
<10-K>
[PASTE OR ATTACH]
</10-K>
Workshop & Facilitation
15. The workshop design prompt
A 3-hour client workshop with 12 stakeholders. This drafts the agenda and the exercises.
Design a [LENGTH]-hour workshop for [N] stakeholders on
[TOPIC].
<context>
Goal of the workshop (one sentence): [STATE IT]
Pre-existing tension among the group: [DESCRIBE]
What "success" looks like at the end: [STATE IT]
Constraints (in person/remote, room layout, prep we can
give): [NOTE]
</context>
Output:
1. Agenda with timing (5-10 minute blocks)
2. One opening exercise that surfaces the tension safely
3. Two working exercises (each: instructions, materials,
facilitator notes on what to watch for, what bad looks
like)
4. The decision/synthesis moment — how we close
5. The risks (which exercise will go off the rails first,
and how we recover)
Avoid icebreakers that adults will hate. No "two truths and
a lie."
16. The pre-read for an executive interview
Before you walk into a 30-minute slot with a CEO.
I have a 30-minute interview with [TITLE] of [COMPANY]
tomorrow. The engagement context is [BRIEF].
<background>
What I know about this person: [LINKEDIN HIGHLIGHTS, RECENT
QUOTES, KNOWN POSITIONS]
What I think their bias on this topic is: [READ]
What I want to learn from this interview: [TOP 3]
What I want them to remember about me: [ONE THING]
</background>
Build me:
1. A 5-question interview guide, ordered to warm up before
getting to the contested questions
2. The one question they probably won't want to answer,
and how I'd phrase it diplomatically
3. The thing I should NOT do (e.g., "don't reference his
predecessor's strategy")
4. A 2-sentence opening that establishes credibility
without taking up the first 5 minutes
Plain-English Translation
17. The "translate this for the operator" prompt
For the moments when the strategy slide has to make sense to a plant manager who doesn't care about the framework.
Below is a section of a deliverable written for a senior
audience. Rewrite it for [OPERATING_ROLE — e.g., "regional
operations director who runs 4 sites"].
<original>
[PASTE]
</original>
Constraints:
- Same content. Same recommendation.
- No consulting jargon. No "leverage," "best-in-class,"
"right-size."
- Concrete examples instead of abstract principles where
possible
- Active verbs, short sentences
- The "why this matters to my Tuesday" angle has to be
visible
18. The "what would a critic say?" prompt
Before you ship anything important, this one stress-tests it.
Below is [DELIVERABLE TYPE — e.g., "the recommendation
section of our final report"]. Take the role of an
intelligent, skeptical reader. The reader is the [CLIENT
ROLE — e.g., CFO] who has to defend this internally.
<deliverable>
[PASTE]
</deliverable>
Tell me:
1. The 3 places the argument is weakest. Be specific.
Quote the sentence and explain the weakness.
2. The 1 place the recommendation is most likely to be
attacked in the room
3. The objection I have not addressed at all
4. The single edit that would most strengthen the document
Be more critical than polite. I'd rather hear it now than
on Wednesday.
What Claude Cannot Do for You
Three honest limits.
It cannot replace the relationship. The trust that closes the deal, the read of the room when the CFO goes quiet, the side conversation before the partner walks in — those are yours. Claude makes you faster between those moments. It does not make those moments.
It cannot decide the recommendation. It can structure your thinking, stress-test it, write it well. It cannot tell you whether to recommend the cost-out or the M&A play. The judgment is the value you charge for. Outsource the typing, not the call.
It will sometimes be confidently wrong. Especially on industry data, dates, and recent events. Cite primary sources for anything that goes into a deliverable. Treat its output like a smart junior's draft — useful, cite-able after you've checked it, never sign-off-able as is.
FAQ
Why Claude over ChatGPT for consulting work specifically?
Three reasons. The 200K-token context window means you can paste a full 10-K, four interview transcripts, and three internal memos into one prompt without summarizing first — and the synthesis quality is meaningfully better than what you get from chunked summaries. Second, Claude follows XML-tagged structured prompts (the kind every prompt above uses) more consistently. Third, in side-by-side use, Claude tends to push back on weak framing rather than just executing — closer to a junior consultant than to a search engine.
Is it safe to put client material into Claude?
Use Claude for Work or Claude Enterprise — those have zero data retention and don't train on your inputs. The free tier of any AI tool is not compliant with most consulting NDAs. Check with your firm's risk team if you're unsure.
How do I get the firm voice into Claude's output?
Two moves. First, paste 3-4 examples of your firm's published memos or proposals into the chat with the instruction "study the voice in these examples and match it." Second, build your "consultant mode" preamble (the one earlier in this article) with firm-specific language conventions — what words you don't use, what structures you favor.
Can I use these prompts with the free version of Claude?
Yes for non-confidential work — practice scenarios, training, prospect research from public sources. No for client-confidential work — that should always go through Claude for Work or your firm's licensed enterprise version.
Where to Go Next
If you're building a consulting AI workflow, the natural next layers are research-synthesis prompts (deeper than what's covered here), financial-model commentary prompts (translating numbers into narrative), and client-onboarding prompt packs (the standard Week 1 deliverables — kickoff agenda, stakeholder map, working norms — automated).
For more individual prompts and specialized templates, the Sales & Outreach and Business Automation categories on SanePrompts include several adjacent prompt packs that pair well with the consulting workflow above. The Claude Prompts for Grant Writing Nonprofits post covers the proposal-writing logic in depth and is a useful pairing for anyone doing pro bono work.
Paste the consultant-mode preamble into Claude after you finish reading. Add your first prompt of the week. The first deliverable you write with this setup will tell you everything about whether it stays in your workflow.