Claude Skills for Consulting Firms: What Clients Can't Copy

Once your clients can load a consulting framework into the same assistant your consultants use, the framework stops being what you sell – what is left is your firm's own method as your best people run it, with evidence the client can check, applied the same way on every engagement. Plenty of clients already pay for those assistants – Microsoft reported "over 30 million paid Microsoft 365 Copilot seats" in July 2026 (Microsoft) – and some Claude skills for consulting firms can be downloaded for free.

My take is that this is good news for consultancies that know what their method is. A generic skill makes one consultant faster. It does not make ten consultants ask the same questions, trust the same sources and stop at the same judgment points – an encoded method does that for the whole firm, and it gets better with every engagement.

Claude skills for consulting firms: what is already public

A skill is one of the "folders of instructions, scripts, and resources that Claude loads dynamically to improve performance on specialized tasks" (GitHub). Two examples are a click away:

• Anthropic's own. The document skills behind Claude's Word, PDF, PowerPoint and Excel work sit in the same public repository, source-available rather than open source, and in October 2025 Anthropic added six finance skills in preview, from comparable company analysis to initiating coverage reports (Anthropic).
• Strategy packs. The slide-software company Oria gives away "21 standalone Claude skills for strategy work" and another set for problem definition, issue trees, hypothesis design, pyramid storylines and objection handling (linked in the FAQ below).

Oria says it plainly: "None of that is proprietary, so it is published here for free." Honest – and exactly why a downloadable framework isn't what a client pays your firm for.

Public skills are good at applying a known framework to a fresh question, at first drafts and at quick desk research. We use skills inside FifthRow ourselves, for decks and support tickets.

In our experience they also break in the same places agents break: they hallucinate, one skill rarely covers a whole job, and quality varies from one run to the next.

Why that turns frameworks into a commodity

If anyone can run a McKinsey-style market entry skill, the framework is no longer what sets your firm apart. When someone tells me a client will just build it with Claude skills, my answer is: "We're still buying furniture, but 3D printers exist." Clients keep buying. They stop paying for the framework on its own.

Partners tell me they feel a squeeze from both sides: clients push fees down because they know you use AI, and the AI bill keeps going up. The way I see it, consulting has three layers – the data, the analysis and the client relationship – and the analytical layer, the part most of us billed the most hours for, is going away. The business model has to change.

What clients still pay for

When a client hires your firm, they are buying the way your best people work:

• the questions your seniors ask first, and the ones they never skip
• the sources they trust in your sector, and the ones they know to ignore
• the order of the steps – what has to be true before the next one is worth doing
• what "good" looks like before a partner puts their name on a draft
• where judgment sits, so the system stops and a person decides
• evidence the client can check, with every claim traced to a source someone can open
• continuity, so the tenth engagement runs on a method the first nine refined

A partner at a brand consultancy told me their edge is relationships and orchestrating the best vendors, not a method. I'd call that a method – which vendor, which brief, what goes back to the client – it just isn't written down yet. Once it is, and it runs as a system, everyone on the team works the way your best people do.

Public skill pack vs your firm's own method

Running your method inside the tools your team uses

FifthRow's Large Consulting Model (LCM) is "a consistent, repeatable, fact-checked way to run consulting work" (FifthRow). Your method runs as Apps inside Copilot, ChatGPT and Claude, so the consultant keeps working where they already work → the firm's method runs behind it. As I put it on a partner call, we take "whatever playbook there is" and "transform it into agents that consistently do that job". Most firms arrive with their own process, so we adapt what we have into "a custom-made thing" – nobody has to code it.

Before anyone on your team reads a draft, Answers strips out made-up citations, and "Quality control is a judge agent in FifthRow that reads every output" (FifthRow). On data, the line is: "We do not train on, resell, or package your data." You keep the client and the last mile. The step-by-step version is in our guide on how to productize consulting services with AI.

What it looks like at a brand strategy boutique

A boutique global brand strategy consultancy works from a four-lens discovery framework its strategists apply to every engagement. Its own leads put that research at 70 percent of the work on a project, and the firm has no junior research bench.

FifthRow encoded the firm's own methodology as running systems. Given only a live engagement's brief and no guidance, the discovery system returned the same lead territory the strategists had already built by hand and presented to their client – in 48 minutes, with 597 verified questions answered. The same framework has since been reused across ten engagements and pitches.

Frequently asked questions

What are Claude skills?

Folders of instructions, scripts and resources that Claude loads when a task needs them. Anthropic ships its own, and anyone can write one or download a public pack, such as Oria's strategy skills.

Can Claude skills replace a consulting team?

They can take over parts of the work: applying a standard framework, a first draft, quick desk research. They don't bring a firm's own questions, trusted sources, quality bar or judgment, and unless something in the setup checks their citations, someone has to do it by hand.

If clients have the same AI tools, why hire a consultancy?

For what AI can't download: the way a firm's best people frame the question, choose the evidence and make the call, applied the same way on every engagement and backed by sources they can check.

What's the difference between a public skill pack and a firm's own method?

A public pack starts as the same instructions for everyone who downloads it, and it runs however each person prompts it. A firm's own method runs as a multi-agent App with the firm's steps, trusted sources and decision points, and every output is fact-checked and reviewed by a judge agent before a consultant sees it.

Two ways to work with FifthRow

You can start with a license for your own delivery – you encode your method and run it, and you keep the client and the last mile. Or you partner in delivery and co-own the work with FifthRow, with clear ownership of the client relationship and the quality bar. Both are on our page for consultancies.

If you want to see one of your own frameworks run, book a slot with me.