For Consulting and Professional Services
How Consulting Firms Can Use AI Without Losing Their Competitive Edge
Your clients can now run ChatGPT or Claude on their own data and get analysis that looks like consultant work. The question is not whether to use AI tools but how to use them so you stay ahead of what clients can do themselves. Without a deliberate approach, AI becomes the thing you present to clients rather than the thing that helps you think better than your competitors.
These are suggestions. Your situation will differ. Use what is useful.
Use AI for speed, not for thinking
Your margin depends on taking client problems seriously enough to build your own view before you synthesise. Copilot and Claude are good at producing plausible summaries of what exists already, which is exactly what clients will ask their tools to do. The point where you add value is where you challenge the consensus or spot what the obvious analysis misses. Use these tools to compress the foundational work so you have more time to form a contrary view, not to replace the time you spend forming any view at all.
- ›Use ChatGPT to summarise financial data or regulatory context so you save three days of junior work, then spend those three days with the client testing whether the consensus diagnosis actually fits their situation
- ›Run a client's own documents through Perplexity to get the baseline landscape, then explicitly ask yourself what questions the baseline analysis did not ask
- ›When you use Claude for initial research, set a rule that you must disagree with at least one thing it said before you take it to the client
Protect the moment where you form independent judgement
The analytical building blocks that used to distinguish junior consultants from each other are now free. What clients pay for is the ability to make sense of those blocks in a way that serves their specific context. If you use AI tools the same way your competitors do, you will reach the same conclusions at the same time. The difference between a consultant and a very expensive report is the moment where you commit to a view that is not the default.
- ›After running analysis through Notion AI or Claude, close the tool and write down what you actually think before you read what the AI suggested
- ›When your team uses Copilot for preliminary work, require them to write their own hypothesis about what is really happening before they search for supporting evidence
- ›In client meetings, name the moment where your view differs from what a standard analysis would show, so the client sees what they are paying for
Change what you charge for as these tools spread
If your fee is pitched as analysis of available data, clients will eventually notice they can buy analysis cheaper. Your client relationships survive because you help them think about what to do with what they know, or because you see a decision differently than they do. As AI commoditises the analytical layer, the value shifts to challenge and to counsel. A firm that still charges the same way as it did before these tools existed will find its margins collapse faster than it expects.
- ›Move your pitch toward the recommendation or the decision framework rather than the analysis that supports it
- ›Charge more for work where you have to form a view against the client's instinct or against what the data seems to show
- ›Offer retainer relationships where you stay close enough to tell the client when things change or when their earlier hypothesis no longer holds
Build practice safeguards that prevent AI from becoming your thinking
The risk is not that AI is too powerful. The risk is that it is convenient enough that your team stops arguing about what is true and starts optimising what sounds credible. This happens gradually and usually without anyone noticing. If your people use Claude or Copilot as their last step before they present, they will begin to defer to it earlier. If they use it as their first step, they will never develop the instinct to know when it is wrong.
- ›Require that initial hypotheses come from your people before any AI tool enters the process
- ›In team discussions, debate alternative explanations for client data before you let anyone run it through Perplexity or ChatGPT
- ›Review how your junior consultants spend their time. If AI has freed them from analytical work but they are not spending that time learning to form judgement, you have a practice problem
Keep your relationships independent from the consensus
Clients hire you because they trust your view of their situation will be different from what they can find elsewhere. The moment all consulting firms start reaching the same conclusions because they use the same tools, client choice becomes about price and speed. This is the moment your relationships become fragile. The only durable defence is a reputation for seeing things differently, which means you have to actually see things differently, not just present AI analysis that sounds more confident than the client's internal view.
- ›After you run analysis through your tools, ask yourself whether your conclusion would surprise your client or whether it confirms what they already suspect
- ›Maintain relationships with people outside your firm who think differently about your sector so you stay exposed to views that AI tools will not generate
- ›When you do reach the same conclusion as a standard analysis would reach, own it explicitly rather than trying to present it as unique insight
Key principles
- 1.AI tools should compress the foundational work that used to occupy time without developing judgement, so your people can spend more time forming views that clients cannot generate themselves.
- 2.The moment you use AI the same way your competitors do, you reach the same conclusions at the same time, and your fees stop being justified by anything but speed.
- 3.Your fee structure should move away from charging for analysis and toward charging for the recommendation or decision framework where your view differs from what the data or consensus suggests.
- 4.If your people use AI tools as their last step before they present, they will begin to defer to those tools earlier in their thinking, and your practice will start to present AI outputs rather than independent counsel.
- 5.Your client relationships depend on a reputation for seeing things differently, which means you have to actually see things differently, not just present confident analysis that sounds different because a better writing tool assembled it.
Key reminders
- Use Copilot or Claude to handle the literature review and data summary that used to take junior consultants a week, then require that your team spend the time they save actually thinking about what it means for this specific client
- Before your team reads what an AI tool suggested, have them write down what they think is actually happening, so the tool becomes a check on their reasoning rather than a substitute for it
- In client conversations, explicitly name the moment where your view differs from what a standard analysis of their data would show, so they see what they are paying for beyond faster access to conclusions they could reach themselves
- Review your fee model annually to see whether you are still charging as though analysis is scarce, and shift your pricing toward the consulting that only happens when someone has formed an independent view
- Build a practice rule that initial hypotheses must come from your people before any AI enters the process, and check this rule quarterly to make sure it is not gradually eroding in the pressure to move faster