Industry Guide: Consulting
How to Use Claude in Consulting: A Guide for Research, Deliverables, and Diligence
Faster research and drafts, same partner-level judgment.
This guide is for consultants, engagement managers, research analysts, and firm leaders who want to put Anthropic's Claude to work on the research, analysis, and deliverables that make up a client engagement. It covers Anthropic's partnership announcements with consulting and professional-services firms and a dedicated customer case study, so you can see concretely how firms are putting Claude to work, not just in theory.
How to use Claude in consulting starts with document-heavy, research-heavy work that has a clear sourcing trail: market research and competitive teardowns built from public filings, due diligence synthesis from data room documents, first drafts of client decks and memos built from research notes, proposal and RFP responses drafted from past capability statements, and knowledge search across a firm's own past engagements. Claude drafts and organizes the material. Consultants and partners still own every client-facing judgment, source every claim back to a named document, and run every deliverable through the firm's normal quality and partner review before it reaches a client.
Published
Free lesson · no account neededContext Is EverythingStrategic Prompting5 minStart the free lessonKey Takeaways
- The strongest consulting uses are market research, client deliverables, due diligence synthesis, proposal drafting, and searching a firm's own past engagement knowledge.
- Anthropic has announced multi-year alliances with Accenture, DXC, and KPMG to roll Claude out across their own consulting and professional-services workforces.
- DXC built an insurance orchestration platform called Assure on Claude, and a sourced customer story describes faster document handling and less manual calculation work.
- Every AI-drafted client deliverable should trace back to a named source document and pass the firm's normal partner and quality review before reaching a client.
- Confidential client data, deal information, and unreleased financials need approved deployments, strict access controls, and a clear confidentiality boundary.
Copy and Paste
Prompts You Can Use Today
Market sizing synthesis from source documents
You are a research analyst on a consulting engagement. Using only the market reports and data tables pasted below, draft a market sizing section that states the total addressable market, the methodology each source used, and the year each figure applies to. Cite the source document by name after every figure, never blend numbers from different years without noting it, and list anything you could not find in the sources instead of estimating it. Source documents: [paste].
Naming the source document after every figure is what makes this fast to check.
Client deck section draft from research notes
Draft the industry trends slide for a client deck using only the research notes and interview summaries pasted here. Write three or four bullet points per section, attribute each insight to the specific note or interview it came from, flag any claim that appears in only one source, and use the firm's standard deck language rather than casual phrasing. Research notes and interview summaries: [paste].
A partner should still review tone and framing before this reaches a client.
Due diligence red flag summary from data room documents
Review the data room documents pasted below and list every item that looks like a possible red flag for this acquisition, such as inconsistent figures between documents, unusual contract terms, customer concentration, or pending litigation. For each item, name the exact document and the section it came from, and do not draw a conclusion about deal risk yourself, only flag it for the deal team to assess. Documents: [paste].
Keeping the risk judgment with the deal team is what keeps this defensible.
Competitive landscape teardown from public filings
Build a competitor comparison table from the public filings, earnings call transcripts, and press releases pasted below. Include revenue growth, stated strategy, and any figures each company has disclosed about the market our client competes in. Cite the filing or transcript by name and date for every number, and mark any competitor where you could not find comparable disclosed data instead of guessing at it. Materials: [paste].
Marking missing data instead of filling it in is what keeps a comparison table trustworthy.
RFP response draft from past proposals
Draft a response to the RFP scope section pasted below using only our past proposals and capability statements pasted here. Match the structure the RFP requests, reuse language from our approved past proposals where it genuinely fits, note which past proposal each reused section came from, and flag any requirement our past work does not clearly address. RFP scope and past materials: [paste].
Flagging weak coverage early gives the proposal lead time to fix it before submission.
From Anthropic
What Anthropic offers for Consulting
Accenture and Anthropic partnership
A multi-year partnership in which Anthropic and Accenture created a dedicated Accenture Anthropic Business Group, a joint offering for chief information officers that pairs Claude Code with Accenture's own adoption methodology, planned offerings for regulated industries, and a co-invested center inside Accenture focused on Claude. Anthropic said the plan trains approximately 30,000 Accenture professionals on Claude and gives tens of thousands of Accenture developers access to Claude Code.
· Source
DXC and Anthropic alliance
A multi-year alliance bringing Claude into systems used by banks, airlines, insurers, and government agencies that rely on DXC's technology. DXC built its internal orchestration platform, called OASIS, with Claude as the default model, and more than 95% of OASIS's code is generated by Claude. Anthropic named DXC part of its network of consulting and systems-integration partners. DXC has 115,000 employees across 70 countries, and over 50 DXC customers are already part of the joint work.
· Source
In Practice
How Consulting Teams Use Claude
According to Anthropic's published customer story
DXC
In a dedicated customer story, DXC describes building Assure, an insurance-industry orchestration platform, on Claude: document intelligence that reads and classifies incoming claims and underwriting documents, workflow orchestration that routes decisions to the right people, purpose-built task apps, and compliance logging that records every action. DXC's insurance business employs 14,000 people and serves 1,100 customers. One team stood up a working workers' compensation benefits calculation app in 8 hours that was 80% right on its first pass, and the share of that calculation requiring human judgment dropped from 70% to 20%.
According to Anthropic's published customer story
Accenture
In its partnership announcement with Accenture, Anthropic said the plan trains approximately 30,000 Accenture professionals on Claude and gives tens of thousands of Accenture developers access to Claude Code, inside a new Accenture Anthropic Business Group built jointly with Anthropic.
According to Anthropic's published customer story
KPMG
In its global alliance announcement with KPMG, Anthropic said the plan reaches KPMG's more than 276,000 employees across 138 countries and territories, through tools KPMG calls Digital Gateway and KPMG Blaze.
Illustrative walkthrough
A market sizing memo for a new client pitch
- A research analyst loads the client's industry briefing, the firm's research template, and the specific market reports approved for this pitch.
- Claude drafts a market sizing section that cites each figure back to the specific report and flags any estimate it could not source.
- The engagement manager checks every cited figure against the original reports and resolves anything flagged.
- A partner reviews the finished section before it goes into the pitch deck.
- The section moves into the pitch through the firm's normal review process.
Illustrative walkthrough
A due diligence data room summary
- A case team uploads the data room's financial statements and management presentations into a secure, scoped workspace.
- Claude extracts the key figures from the cited documents only, flagging gaps instead of filling them in.
- An associate checks the extraction against the source documents and resolves any flagged gap with the deal team.
- A partner reviews the summary and decides what belongs in the diligence report.
- The summary moves into the diligence report through the firm's normal review.
Core Workflows
Claude Consulting Use Cases From Research to Delivery
These are the areas where AI for consulting firms saves the most analyst and associate time while keeping a sourcing trail a partner can check.
Market research
AI for market research turns public filings, industry reports, and transcripts into a sized market and a sourced competitive picture.
Client deliverables
AI for client deliverables drafts deck sections and memos from research notes, in the firm's own language and structure.
Due diligence synthesis
Data room documents become a structured summary with every figure and flag traced back to a specific document.
Internal knowledge management
Search across a firm's own past engagement materials to reuse frameworks and findings instead of starting from a blank page.
Proposal and RFP drafting
First drafts of proposal and RFP responses built from approved past proposals and capability statements, with gaps flagged for the proposal lead.
Workshop and stakeholder prep
Workshop agendas, stakeholder interview guides, and synthesis of interview notes into themes for a steering committee.
What Consulting Firms Report
How Consulting Firms Use AI Today
How is AI used in consulting firms right now? Most of what is public is not an independent customer case study, it is Anthropic naming a consulting or professional-services firm as an implementation partner that is also rolling Claude out across its own workforce. That framing matters: these are partnership and alliance announcements, not third-party reviews, and they are worth reading as such.
Anthropic and Accenture announced a multi-year partnership that created a dedicated Accenture Anthropic Business Group, a joint offering for chief information officers that pairs Claude Code with Accenture's own adoption methodology, planned offerings for regulated industries, and a co-invested center inside Accenture focused on Claude. Anthropic said the plan trains approximately 30,000 Accenture professionals on Claude and gives tens of thousands of Accenture developers access to Claude Code.
Anthropic and KPMG announced a global alliance that reaches KPMG's more than 276,000 employees across 138 countries and territories, through tools KPMG calls Digital Gateway and KPMG Blaze. Here too, the headline figures describe KPMG's own workforce rollout, not an outside audit of results.
DXC and Anthropic announced a multi-year alliance bringing Claude into systems used by banks, airlines, insurers, and government agencies that rely on DXC's technology, across a company of 115,000 employees in 70 countries. DXC is the one name in this group with both an alliance announcement and a separate, dedicated customer story with concrete figures, which is why it anchors the case study section below.
Anthropic's Alliances
The Claude Partnerships Reshaping Professional Services
Generative AI in consulting is moving fastest through firm-level partnerships rather than individual point tools. Accenture, DXC, and KPMG are each building their own internal platforms and methodologies around Claude, then offering that work to their own clients as a service.
DXC built its internal orchestration platform, called OASIS, with Claude as the default model, and Anthropic said more than 95% of OASIS's code is generated by Claude. Anthropic named DXC part of its network of consulting and systems-integration partners, and said over 50 DXC customers are already part of the joint work.
For an individual consulting team, the practical takeaway is less about any single branded platform and more about the pattern underneath it: narrow, well-scoped AI workflows wrapped in a firm's own methodology, review steps, and client-facing judgment, rather than an AI tool working unsupervised on a client's behalf.
If your firm is not yet part of one of these larger alliances, the same pattern still applies at team scale: pick a narrow, repeatable workflow, wrap it in your own templates and review steps, and measure it on your own work before expanding it.
Research in Practice
AI for Strategy Consulting Done Well
The pattern that holds up in market research and strategy consulting is narrow scope and hard citations. Give Claude the specific reports, filings, or transcripts it may use, your deck template, and your house style. Ask it to cite the source document after every figure and to mark anything it cannot find instead of filling the gap. A draft built that way can be checked source by source, which is what a case team actually needs before a partner sees it.
Cross-document consistency checking is an underrated use of AI for client deliverables. A long engagement produces dozens of working documents, and a number that drifts between a model, a memo, and a deck slide is exactly the kind of error a client asks about in the room. Claude can compare documents and list every place they disagree, turning a full reread into a short, focused list.
Keep the division of labor explicit. Claude drafts, organizes, and cross-checks. Analysts and associates own accuracy and clarity, engagement managers own structure and story, and partners own what actually goes in front of a client and how it is framed. Writing that division into how your team works makes AI-assisted drafting easier to defend when a client asks how a number was produced.
Choosing Tools
Claude vs ChatGPT for Consulting
The Claude vs ChatGPT for consulting comparison should run on your own work: a market sizing section already delivered, a due diligence summary already checked against the data room, and a proposal section already submitted. Compare accuracy against the original sources, how clearly each tool cites what it used, and how each one handles a gap in the source material.
In consulting the decision is usually made on confidentiality and workflow fit rather than on a feature checklist: which deployment your IT and risk teams approve for client data, how long documents are retained, whether the tool can be scoped to one engagement's materials, and how well it fits the deck, document, and spreadsheet formats your firm already uses.
Ask the same question about any AI tool you are evaluating: can it be restricted to only the documents I give it, and can I see exactly which document a claim came from. A tool that cannot answer both questions clearly is a weaker fit for client-facing work, regardless of which model sits behind it.
Getting Started
A Safe First Month for a Consulting Team
Start with a closed engagement
Test prompts on a deliverable from a finished, closed engagement, and compare the AI draft with what actually shipped to the client.
Load templates and style guides
Put your deck templates, house style, and glossary into a project so drafts match what a partner expects to see.
Choose one deliverable type
Pick one recurring deliverable, such as market sizing sections or proposal drafts, and measure review time against the usual baseline.
Define the confidentiality boundary
Agree with risk and IT which client data may touch which deployment, and how AI-assisted drafts are reviewed and stored.
Limits
Where Consultants and Partners Must Decide
Can AI replace consultants? No. Claude can misstate a figure, miss the context behind a number, or draft a conclusion the underlying research does not actually support. A client is paying for judgment about what a finding means for their specific situation, a relationship built on trust, and accountability for the recommendation, none of which a model can carry.
Client-facing judgment calls, including what a research finding actually means for a client's strategy, how to frame a sensitive finding in the room, and what belongs in a final recommendation, stay with consultants and partners. Claude prepares material for those judgments and does not make them.
Confidential client data, deal information, unreleased financials, and anything covered by a non-disclosure agreement are highly sensitive. Use only approved deployments, scope access to the specific engagement, and keep retention and deletion aligned with what you agreed to with the client.
Build Reliable AI Habits for Client Work
Getting trustworthy drafts from AI is a practiced skill: sourcing, citation, and review. The free Nightschool AI curriculum is hands-on from the first lesson.
Frequently Asked Questions
How is AI used in consulting firms?
Mainly for research, drafting, and knowledge search: market sizing and competitive teardowns, first drafts of client decks and memos, due diligence synthesis, and proposal and RFP drafting. Anthropic has announced alliances with Accenture, DXC, and KPMG to roll Claude out across their own consulting workforces, and these are mostly partnership announcements rather than independent reviews. Client-facing deliverables still go through the firm's normal partner review.
Can AI help with client deliverables?
Yes, for first drafts: deck sections, memos, and summaries built from research notes and source documents you give it, written in your firm's house style. It works best with a narrow scope and a citation after every figure. A partner should still review the finished deliverable before a client sees it.
How do consulting firms deploy AI safely?
By scoping access to one engagement's materials, agreeing with risk and IT which client data may touch which deployment, and keeping retention and deletion aligned with client agreements and non-disclosure terms. Starting on a closed, finished engagement before using AI on live client work is a practical way to build confidence.
What are the biggest AI use cases in consulting?
The most common are market research and competitive analysis, client deliverable drafting, due diligence document synthesis, proposal and RFP drafting, and searching a firm's own past engagement knowledge. Each works best with named source documents and normal partner review.
Can AI replace consultants?
No. AI can draft and organize research, but a consultant's judgment about what a finding means for a specific client, the relationship built with that client, and accountability for the recommendation are not things a model carries. AI assists the drafting and research; the recommendation stays with the consultant.
How much time can AI save on consulting work?
It varies with the document type and how clean the source material is, so measure it on your own deliverables. Time savings come mainly from faster first drafts of research sections and summaries built from sources you already have. Review and partner sign-off time does not disappear, and it should not.
Is AI accurate enough for client-facing deliverables?
Only when it is grounded in the specific documents you give it and every claim is checked against those sources before a client sees it. An AI draft built without hard citations to named source documents is not reliable enough for client-facing work on its own.
Keep going
Related Guides and Courses
Related reading
- AI Training for Work
- Free AI Prompt Generator for Work Tasks
- How to Use Claude in Legal
- How to Use Claude in Accounting
- How Consultants Use Claude
- Claude for Executives
- How Researchers Use Claude
- How to Use Claude in Real Estate
- How to Use Claude in Education
- How to Use Claude in Marketing
- Private Equity Due Diligence With Claude: Data Rooms to Findings