The discovery call is where deals are won or lost. Not when the proposal is sent—that is just documentation. The actual buying decision happens during the conversation where the prospect realizes that you understand their problem better than they do.
Most AI agency discovery calls are casual conversations that meander through vague topics and end with "let me send you a proposal." The prospect leaves unclear about your approach, and you leave without enough information to write a good proposal. Both sides are winging it.
A structured discovery framework changes everything. It ensures you gather the information you need, demonstrate your expertise, build trust, and create urgency—all in forty-five to sixty minutes.
Pre-Call Preparation
The discovery call starts before you join the video. Fifteen minutes of research dramatically improves the quality of the conversation.
Research Checklist
- Company website: Understand their products, services, and market
- LinkedIn profiles: Review the backgrounds of everyone attending the call
- Recent news: Any press releases, funding rounds, product launches, or leadership changes
- Job postings: What are they hiring for? AI-related roles signal investment. Operations roles signal scaling pain.
- Industry context: What challenges is their industry facing right now?
- Competitor intelligence: What are their competitors doing with AI?
Hypothesis Development
Based on your research, develop two to three hypotheses about their situation:
- "Based on their growth and the operations roles they are hiring for, they are probably struggling with manual processes that do not scale"
- "Their industry is facing new compliance requirements, which may be driving interest in AI governance"
- "Their competitor just launched an AI-powered feature, which may be creating competitive pressure"
You will test these hypotheses during the call. Having them prepared makes your questions more targeted and your insights more impressive.
The Call Structure
Opening (5 minutes)
Goal: Set the tone, build rapport, and establish the agenda.
Start with a brief personal connection (thirty seconds, no more), then set the agenda:
"Thanks for making time. I want to make sure this conversation is valuable for you regardless of whether we work together. Here is what I would like to cover: I want to understand your current situation and the specific challenges you are facing. Then I will share some relevant experience and, if it makes sense, we can discuss potential approaches. Does that work?"
This framing accomplishes three things:
- It positions you as a consultant, not a salesperson
- It sets expectations for a structured conversation
- It signals that you might not be the right fit (which paradoxically builds trust)
Problem Exploration (15-20 minutes)
Goal: Understand the prospect's situation deeply enough to design a solution.
Follow this question sequence:
Current State Questions:
- "Walk me through how [relevant process] works today. Start from the beginning."
- "Where does it typically break down or slow down?"
- "How many people are involved in this process? What are their roles?"
- "What tools or systems are you currently using?"
History Questions:
- "How long has this been a challenge?"
- "What has changed recently to make this a priority now?"
- "Have you tried to solve this before? What happened?"
Impact Questions:
- "How is this affecting the business? Can you put a number on it?"
- "How much time does your team spend on this weekly?"
- "What is the error rate, and what happens when errors occur?"
- "If this problem gets worse over the next twelve months, what does that look like?"
Insight Sharing (5-10 minutes)
Goal: Demonstrate expertise and build credibility by connecting their situation to your experience.
After you understand their problem, share relevant insights:
"What you are describing is very similar to what we see in other [industry] companies at your stage. The pattern is usually [describe the pattern]. What typically works is [describe the general approach]. For example, one of our clients in [similar industry] was dealing with the same bottleneck, and we [brief description of approach and outcome]."
This is not a pitch. It is a demonstration that you have solved this before and understand the nuances. The prospect should feel that you "get it."
Decision Process Exploration (10 minutes)
Goal: Understand how they will make a decision so you can align your proposal.
- "Who else needs to be involved in this decision?"
- "Is there a budget allocated for this, or does budget need to be created?"
- "What does your evaluation process look like? Are you talking to other vendors?"
- "What criteria will you use to make the decision?"
- "What would prevent this from moving forward?"
- "What is your ideal timeline for having a solution in place?"
These questions are critical. Most agencies skip them and then are surprised when the proposal stalls in procurement for three months or gets killed by a stakeholder they never met.
Technical Feasibility Check (5-10 minutes)
For AI projects specifically, you need to assess whether the problem is technically solvable with their current data and infrastructure.
- "What data do you have related to this process? Is it structured or unstructured?"
- "How is the data currently stored and accessed?"
- "Are there any regulatory or compliance requirements we should be aware of?"
- "What is your current technical infrastructure? Cloud provider, key systems?"
- "Are there any integration requirements with existing tools?"
This section serves a dual purpose: gathering technical information and demonstrating that you think carefully about feasibility before proposing a solution.
Next Steps (5 minutes)
Goal: Secure a specific commitment for the next step.
Summarize what you heard:
"Let me make sure I have this right. You are dealing with [problem], it is costing you approximately [impact], you have tried [previous attempts], and you need a solution by [timeline]. The decision involves [stakeholders] and the budget is [range]. Did I miss anything?"
Then propose the next step:
"Based on what you have shared, I believe we can help. Here is what I suggest as a next step: I will put together a proposal that includes our recommended approach, a timeline, and investment options. I can have that to you by [date]. Then let us schedule a follow-up to walk through it together. Does [specific date] work?"
Always schedule the follow-up before ending the call. "I'll send you a proposal" without a follow-up date is a recipe for deals that go dark.
Advanced Discovery Techniques
The "What Would Success Look Like?" Question
Ask this early: "If we are sitting here in six months and this project was a complete success, what would be different?"
This reveals their true priorities and success criteria, which should drive your proposal design.
The Cost of Inaction
Help the prospect quantify the cost of doing nothing:
"You mentioned this process costs roughly $X per month in labor and errors. If it takes three months to implement a solution, that is $3X in continued costs. And if the problem grows with your volume, the cost next year could be $Y."
This creates urgency without being pushy. You are simply doing the math they have not done.
The Pre-Mortem
Ask: "What would cause this project to fail?"
This reveals hidden concerns, political dynamics, and risk factors that the prospect might not volunteer otherwise. It also shows that you think carefully about risk, which builds trust.
The Champion Test
If you suspect the person on the call is a champion but not the decision-maker, test it:
"If you and I agreed that this was the right approach, what would the next step look like internally? Who else would need to review and approve?"
Their answer tells you whether you are talking to the right person and what the internal buying process looks like.
Post-Call Actions
Within 1 Hour
Send a brief summary email:
"Thanks for the conversation today. Here is my understanding of your situation: [2-3 bullet summary]. Based on this, I will prepare a proposal covering [approach outline] and have it to you by [date]. Our follow-up is scheduled for [date/time]. In the meantime, if any questions come up, do not hesitate to reach out."
Within 24 Hours
- Enter detailed notes in your CRM
- Begin proposal preparation
- Share relevant case studies or resources with the prospect
- Brief your team on the opportunity
Call Scoring
After every discovery call, score it:
- Problem clarity (1-5): How well do you understand their problem?
- Budget qualification (1-5): How confident are you in their budget?
- Authority confirmation (1-5): Are you talking to the decision-maker?
- Timeline urgency (1-5): Is there a real deadline driving this?
- Technical feasibility (1-5): Can you actually solve this with AI?
- Engagement quality (1-5): How engaged was the prospect?
Deals scoring 24+ out of 30 are strong opportunities. Below 15, consider deprioritizing.
Common Discovery Call Mistakes
- Talking more than listening: The prospect should talk 60-70% of the time. If you are doing most of the talking, you are pitching, not discovering.
- Jumping to the solution too early: Resist the urge to start proposing solutions before you fully understand the problem.
- Skipping the decision process questions: You cannot write a winning proposal if you do not know how the decision will be made.
- Not quantifying the impact: If you cannot put a number on the problem, you cannot build an ROI case.
- Ending without a scheduled next step: "Let me send you a proposal" is not a next step. A calendar invite is a next step.
- Not taking notes: Your proposal will be vague and generic if you do not capture the specific language and numbers from the call.
- Being afraid to disqualify: If the prospect is not a fit, it is better to say so now than to waste weeks on a proposal that will not close.
The discovery call is your most important sales asset. Master it, structure it, and refine it based on what works. A forty-five minute conversation that follows this framework will outsell a polished pitch deck every single time.