The quality of your discovery call determines whether you close the deal. Not the proposal, not the follow-up, not the pitch. The call. And the quality of the call depends almost entirely on the quality of your questions.
Generic questions, “What are your goals?” and “What challenges are you facing?”, produce generic answers. Vague answers lead to proposals that don’t land, misaligned scope, and projects that start without shared understanding of what success means.
Specific questions produce specific answers. The problem is generating specific questions for each different buyer type, without spending an hour on prep for every call. This is exactly what AI is good at.
The Prompt That Generates 12 Targeted Questions
Run this prompt in Claude 30 minutes before every discovery call:
The Full Prompt:
“I’m meeting with a [buyer title] at a [company type] who is experiencing [pain category]. Generate 12 discovery questions for this conversation.
The questions should probe across five areas:
- Root cause (2 questions), what’s driving this problem, not just the surface symptom
- Business impact (2-3 questions), what does this cost them in revenue, time, or capacity
- Current solution (2 questions), what they’ve already tried, why it didn’t work
- Decision process (2-3 questions), who is involved, what does approval look like, what is their timeline
- Success criteria (2 questions), how they will know the engagement worked
Order the questions from least to most probing, start with context-gathering, end with the questions that require trust.
Avoid: yes/no questions, leading questions, questions with ‘always’ or ‘never,’ and generic questions that could apply to any buyer. Every question should be specific to the context I’ve provided.”
What you get: 12 questions ordered by depth, covering the full discovery arc.
What you do next: read all 12, circle the 5-7 most relevant to what you already know about this buyer, add 1-2 of your own based on your experience, and walk into the call with a focused set.
The Five-Category Framework
Understanding why each category matters helps you evaluate AI-generated questions and fill gaps manually.
Category 1: Root Cause Questions
Surface: “We’re spending too much on paid ads.” Root cause: “Our conversion rate from ad traffic is 0.8% when it should be 3-4%, which makes every acquisition channel unprofitable.”
Root cause questions work because they shift the problem from a symptom you might address superficially to the actual problem that requires your real expertise. When buyers articulate their root cause out loud, they often discover the scope of the problem themselves, which is how they become motivated to solve it.
Example AI-generated root cause question: “When you look at the gap between where your marketing spend is and where the results are, which piece of the funnel do you think is most broken, and why do you think that’s been hard to fix internally?”
Category 2: Business Impact Questions
Buyers don’t buy services. They buy outcomes. Impact questions quantify the cost of the current situation and make the business case for change.
Example: “If this challenge continues at the current rate for another 12 months, what does that cost your team, in time, revenue, or both?”
This question does two things: it forces the buyer to calculate the cost of inaction, and it gives you the ROI anchor for your proposal. If the answer is “it costs us $200K in revenue per year,” and your engagement is $25K, the business case writes itself.
Category 3: Current Solution Questions
You need to know what they’ve tried. Not to criticize it, but to avoid proposing the same thing and to understand what the actual barrier is.
Example: “What approaches have you already tried to address this? And when you say they didn’t work, did they produce no results, or partial results that weren’t sustained?”
“No results” and “partial results not sustained” require different solutions. This question prevents you from proposing work that has already been attempted.
The buyer who says “we’ve tried three agencies and none of them worked” is not a red flag. They’re a buyer who has already proved budget and willingness to invest. Your job is to understand why the previous attempts failed, not to be alarmed that they failed.
Category 4: Decision Process Questions
Proposals die because the person you talked to wasn’t the actual decision-maker, or because the timeline was unrealistic, or because budget wasn’t allocated. Decision process questions prevent all three.
Example: “Once you find the right person for this engagement, what does the approval process look like, and who else needs to be involved in that decision?”
Ask this question early, not at the end of the call. If the answer is “there’s a three-month procurement process and a committee of six,” your close timeline and proposal structure need to account for that. Finding out on the follow-up call is too late.
Category 5: Success Criteria Questions
The most important question most solos skip: “How will you know this worked?”
Ask it before you start, not at the end of the project. Buyers who can’t articulate success criteria are not ready to buy. Buyers whose success criteria don’t match what your service actually delivers are not a fit. Surface this in the discovery call, not in the project post-mortem.
Example: “Six months from now, if this engagement exceeded your expectations, what specifically would be different about your business? What number would have moved, or what capability would you have that you don’t have now?”
Building Your Question Bank
After every discovery call, log the questions that worked:
- What question opened up the most useful conversation?
- What question produced a specific, revealing answer?
- What question did the buyer respond to most viscerally?
Over 20 calls, you build a library of proven questions by buyer type. The AI generates a starting set; your call experience refines it.
Store your bank in a simple structure:
/Question Bank
/VP of Marketing at SaaS company
- Root cause questions [5 best]
- Impact questions [5 best]
/Operations Director at agency
- Root cause questions
- Impact questions
/Founder at early-stage startup
...The more specific your question bank, the better your calls. A question designed for a VP of Marketing at a 50-person SaaS company is more effective than a question designed for “buyers.”
The Question Selection Process (5 Minutes Before the Call)
After AI generates the 12 questions, select your 5-7 using this criteria:
- Must ask: The one question that will most reveal whether this is a real fit. (Usually an impact question.)
- Should ask: 3-4 questions that build the picture of the situation. (Root cause + current solution.)
- Will ask if time allows: 1-2 deeper questions for if the conversation goes well. (Success criteria, decision process details.)
Write these on paper, not a screen. Looking down at a notebook signals thinking; looking at a screen signals reading.
The questions are a framework, not a script. Go into the call prepared to follow the conversation wherever the prospect leads. The best question is often a follow-up to what they just said, not the next item on your list.
The best discovery calls don’t feel like discovery calls. They feel like a conversation between two people who both understand the problem deeply. Your questions are the instrument that creates that feeling, they signal that you’ve done your homework and that you’re genuinely trying to understand, not just qualify.
The Recovery Question
Every discovery call has a moment where the conversation stalls or the prospect gives a vague answer. This is your universal recovery question:
“Can you give me a specific example of when that happened?”
Vague answer: “Our marketing results have been inconsistent.” After recovery: “Last Q4, we ran a $40K campaign and generated 12 leads, when we needed 50. The quarter before, we ran a $15K campaign and got 40 leads. We can’t figure out what made the difference.”
That specific answer tells you everything. The recovery question works because examples are always more informative than generalizations, and buyers always have examples, they just default to generalization until you ask.
Memorize this question. It’s worth more than all 12 AI-generated questions combined in a stalled conversation.
The Integration With Your Close Rate
Track your discovery call conversion rate over 30 calls before and after implementing this system.
Baseline: what percentage of discovery calls convert to proposals? What percentage of proposals close?
After 30 calls with prepared, targeted questions: both numbers should move. Close rate improvements from better discovery are typically in the 10-20 percentage point range because:
- You identify poor-fit prospects earlier and invest less in their proposals
- Your proposals address the specific root cause they articulated, not a generic problem
- Your success criteria are pre-aligned, so the buyer says “yes, this is exactly what I need”
The questions are not magic. They’re infrastructure. Install the infrastructure, work the calls, and measure the results.
Ready to send stronger proposals?
Build, send, and track proposals in one place so follow-up is easier.
Start your free trial →
