For Consulting and Professional Services

40 Questions Consulting and Professional Services Should Ask Before Trusting AI

Your fee justification depends on independent insight your clients cannot easily generate themselves. When you run analysis through Copilot or Claude, you need to know what you are actually getting and whether it protects the judgement that clients pay for.

These are suggestions. Use the ones that fit your situation.

Download printable PDF

Questions About What AI Actually Analysed

1 When Claude summarised that market report for your proposal, did it read the full document or did it extrapolate from the title and opening section?
2 The competitive analysis ChatGPT generated uses information from when exactly, and does your client need current data that falls outside that training window?
3 You asked Copilot to spot patterns in your client's sales data. Did it actually examine the full dataset or did it work from the sample you pasted into the chat?
4 When Perplexity returned citations for the regulatory landscape you are advising on, did it pull from primary sources or from secondary articles about those sources?
5 Your junior ran industry benchmarks through Notion AI. Do you know whether it sourced from published databases your client could access directly or from general web content?
6 The AI recommended a restructuring approach based on comparing your client to unnamed peers. How would you defend the relevance of those comparisons in a client conversation?
7 ChatGPT provided three strategic options for your client problem. Did it weight these by likelihood of success, or did it simply generate plausible alternatives without evaluation?
8 You used Claude to stress-test assumptions in your financial model. Which assumptions did it actually test, and did it flag which ones it lacks domain knowledge to challenge?
9 The AI output contradicts a conclusion from your last engagement with this client. Can you explain why the AI reached a different conclusion without blaming it on training data differences?
10 Copilot generated talking points for a client presentation. Does the analysis behind those points exist in your workings, or only in the AI output you have not yet verified?

Questions About Whether Clients Can Replicate Your Work

11 Your analysis used the same public tools and datasets your client could buy themselves. What specific interpretation or insight did you add that justifies your fee?
12 A client asks why your recommendation differs from what ChatGPT told them when they pasted the same brief into it. How do you answer?
13 You are charging for competitive intelligence. Your source was Perplexity, which your client can access for free. What prevents them from simply running the same searches?
14 Your proposal includes a diagnostic analysis that mostly involved running client data through Claude and organising the output. Would a client consider this work they could outsource to a junior analyst with access to the same AI tools?
15 A partner in your firm could generate your exact analysis by spending two hours with Copilot instead of two weeks of junior time. Does the client see any difference in the output?
16 You are advising on operational efficiency. The recommendations came from a systematic analysis using AI. How is your judgement visible in the final advice, rather than the AI system's pattern matching?
17 Your client could theoretically hire a single smart analyst with a Copilot licence to do what your team does now. What structural value does your firm add that prevents this?
18 The strategic insight you are presenting came from asking Claude the right question rather than from months of primary research. Does your client need to know the difference?
19 Your client asks for the raw AI output you worked from, before your interpretation. Would showing them make your added value clearer or less convincing?
20 A client wants to see your working on a recommendation you partly developed through ChatGPT. What documentation do you actually have that shows your thinking, not just the AI's output?

Questions About Judgement and Independence

21 You have been asked to challenge a client's strategy. Did you approach this analysis expecting to find problems, or did you ask Claude to evaluate the strategy and then present its conclusions as your own?
22 The same AI tools your firm uses are also used by your competitors and by clients who consider hiring you. How does this affect whether your recommendations feel independent or like industry consensus?
23 Your team used Notion AI to organise interview notes from client stakeholders. When you now synthesise those organised notes into recommendations, are you bringing your own judgement or validating what the AI already surfaced?
24 You asked ChatGPT to propose three possible approaches to a client problem. You then selected one and presented it as your recommendation. Did you actually reject the others, or did you simply pick the first one you thought was defensible?
25 A junior consultant shows you AI generated analysis that directly contradicts your instinct about a client situation. How do you decide whether the AI has spotted something real or whether your experience trumps the pattern matching?
26 You are tempted to present an AI insight because it is clever and surprising and would impress the client. What question should you ask yourself before doing this?
27 When you stop using AI tools on a project and go back to first principles analysis, how often do you reach different conclusions, and what does that tell you about whether you are thinking independently?
28 Your team has started using Claude to generate options before they discuss them as a group. Has this improved your collective judgement, or does it anchor everyone to the AI's framings before the real debate starts?
29 The client problem involves a decision where there is a clear optimal answer in the published literature. Your job is to recommend it with confidence. What does AI generate do to your credibility in this situation?
30 A partner uses Copilot to sense-check a recommendation you made. The AI agrees with you. Do you feel more confident in your judgement, or have you just validated against a system that works similarly to how you think?

Questions About What You Are Charging For

31 Your day rate has stayed the same since you started using AI tools that reduced the time you spend on foundational analysis. Should it have changed?
32 You are billing a client for analysis time. Your team completed the work in half the expected hours because Copilot accelerated the initial mapping. How do you bill this fairly without signalling that the analysis is now worth half what you originally quoted?
33 A client wants to know what they are paying for when you use ChatGPT in your process. What is the honest answer?
34 Your proposals used to justify fees by promising rigorous analysis that would take significant time. That promise feels weaker now that AI can do the initial legwork quickly. What new promise are you making instead?
35 A junior consultant was developing judgement by wrestling with complex data analysis over months. Now Notion AI handles the initial synthesis. How does this affect what you can credibly charge that junior's time toward?
36 You are competing for a client engagement against a boutique firm that markets itself as offering independent strategic thinking without algorithm bias. How do you respond?
37 Your firm has always charged a premium for access to proprietary data or unique methodologies. If your methodologies now mostly involve asking Claude better questions, does that premium hold?
38 A client suggests they could do this work themselves with AI tools if they had the right training. Is your response about the insights you generate, or the time you save them, or something else?
39 You are renewing a retained advisory relationship. The client sees that your team now uses the same AI tools available to their internal team. Why should they still pay you?
40 Your practice development conversation with a prospect now hinges on whether you can offer insights they cannot generate. Are you winning these conversations the same way you did five years ago?

How to use these questions

Related reads

The Book — Out Now

Cognitive Sovereignty: How To Think For Yourself When AI Thinks For You

Read the first chapter free.

No spam. Unsubscribe anytime.