Client Discovery & Needs Assessment

Discovery is the most important skill in AI consulting and the one most technical founders underinvest in. A good discovery conversation surfaces the real problem (which is often not what the client initially describes), establishes whether there's budget and authority to solve it, and positions you as the expert who understands their situation better than they do. Structure your discovery call in four phases. Phase 1 — Context (10 mins): understand their business, their current processes, and what prompted them to explore AI now. What changed? What's the pain costing them? The goal is to hear their world in their language. Phase 2 — Pain Excavation (15 mins): go deeper on the specific problem. Ask: 'What does that cost you in time or money?' 'What happens if this isn't solved in the next 12 months?' 'How are you handling it today?' The SPIN methodology from Huthwaite is worth studying here — Situation, Problem, Implication, Need-payoff. Phase 3 — Vision (10 mins): help them articulate what success looks like. 'If we solved this perfectly, what would your business look like in six months?' This anchors the conversation on value, not cost. Phase 4 — Next steps (5 mins): never end without a clear next action and timeline. Either you're moving to proposal or you're scheduling a follow-up with a specific decision-maker in the room. For AI consulting specifically: listen for the underlying business problem, not the technology request. A client who says 'we want to build a chatbot' is really saying 'we have a customer service problem' or 'our team is spending too much time answering repetitive questions.' Solve the business problem. The technology is how.

Before scoping any AI engagement, you need to assess the client's readiness. Many businesses that are excited about AI aren't yet ready to execute effectively — their data is a mess, their processes are poorly defined, or their leadership don't actually understand what they're buying. Knowing this early saves both parties significant time and money. The five dimensions of AI readiness: (1) Data quality — Do they have the data the AI system will need? Is it clean, accessible, and in usable formats? Most AI projects underestimate the data preparation work by 40–60%. (2) Process clarity — Can they explain the workflow the AI will support in clear, step-by-step terms? If they can't describe the process to you, you can't automate it. (3) Change management — Is there a champion who will drive adoption internally? AI tools that aren't adopted provide zero ROI. (4) Technical infrastructure — Do they have APIs, cloud infrastructure, or are you starting from spreadsheets and email? (5) Budget clarity — Are they treating this as a genuine investment or fishing for a quote they'll use to benchmark a decision they've already made elsewhere? Build a simple AI readiness scorecard (1–5 on each dimension) that you fill out during or immediately after discovery. This gives you a structured basis for your proposal and often surfaces conversations that need to happen before any work begins. Clients who score low on data or change management need a different engagement than clients who are ready to build — your proposal should reflect that.