Discovery Call Script for an AI Agency (With Timings)

A discovery call script for an AI agency runs about 30 minutes: set the agenda, map the current process, put a number on the problem, check who decides, budget and timing, then agree on a dated next step. Let the client do most of the talking.

Below is the full script, section by section, with the questions to ask. Adapt the wording to your voice and niche. The example niche is a home services company, and the automation being discussed is lead follow-up.

Before the call

Spend ten minutes preparing:

• Look at their website, reviews and booking flow. Test the contact form if you haven't.
• Note what software they appear to use (booking widget, chat tool, CRM clues in their emails).
• Write down two or three guesses about where their process leaks.
• Have your demo video ready to share, but don't plan to show it until the end.

The call structure

Section	Time	Goal
1. Open and agenda	0:00 to 0:03	Set expectations
2. Current process	0:03 to 0:12	Understand how the work happens today
3. Cost of the problem	0:12 to 0:18	Put a number on it
4. Decision, budget, timeline	0:18 to 0:24	Qualify
5. Fit and next step	0:24 to 0:30	Agree what happens next

1. Open and agenda (3 minutes)

Thanks for making the time, [Name].

Here's what I'd suggest for the next 30 minutes. I'll ask some questions about
how leads come in and get handled today, so I understand your setup. Then, if
it looks like something I can help with, I'll tell you how I'd approach it and
we can decide whether a next step makes sense. If it's not a fit, I'll say so.

Does that work for you? Anything you want to make sure we cover?

Asking what they want to cover gives you their real priority early.

2. Current process (9 minutes)

Get them to walk through the process step by step. Write it down as they talk.

Walk me through what happens when a new lead comes in today, from the moment
they fill in the form or call.

Who sees it first? How do they find out it's there?

How long does it usually take before someone gets back to them?

What happens after hours or on weekends?

What software do you use for this? (CRM, calendar, phone system, email.)

Roughly how many leads come in a month? How many through each channel?

Where does it go wrong? Tell me about the last lead that slipped through.

The last question is the most useful one. A specific story ("a homeowner called three times and we never called back") tells you more than any general answer.

3. Cost of the problem (6 minutes)

Now turn the problem into numbers. Do the math with them, out loud.

Of those leads, roughly how many turn into a booked estimate?

And how many estimates turn into jobs?

What's a typical job worth to you, after costs?

If you had to guess, how many leads a month go cold because nobody got back
to them fast enough?

How many hours a week does your team spend chasing and following up?

So if we got back even two of those jobs a month, that's about [$X] a month.
Does that sound right to you?

Let them correct your math. When the client says the number, they believe it. Don't invent a number for them.

4. Decision, budget and timeline (6 minutes)

These questions qualify the deal. Ask them plainly.

Besides you, who else would be involved in deciding on something like this?

Have you tried to fix this before? What happened?

Is there a budget set aside for this, or would it come out of something else?

For context, projects like this usually have a one-time setup fee and a
monthly fee for running and supporting it. Is that structure workable for you?

Is there a reason to get this done by a particular date, like a busy season?

If everything else looked right, what would stop you from going ahead?

How to qualify

By the end of section 4 you should know:

• Problem: is it real, and do they feel it?
• Value: is the monthly cost of the problem big enough to justify your price?
• Authority: are you talking to the decision-maker, or do others need to join?
• Budget: is there money, and does your pricing structure work for them?
• Timing: is there a reason to act now?

If two or more of these are missing, it's usually better to say so and set a later follow-up than to send a proposal that sits unread.

5. Fit and next step (6 minutes)

Summarize, give a short view of how you'd solve it, then agree on a dated next step.

Let me play back what I heard. Leads come in through the website and phone,
about [number] a month. After hours, they go to voicemail, and it's often the
next day before anyone calls back. You think around [number] a month go cold,
and each job is worth about [$X]. Did I get that right?

Here's how I'd approach it: every new lead gets a text and email within two
minutes, with a link to book the estimate. After hours, calls get answered
and the details logged. Your team sees everything in one place.

Here's a 3-minute video of something similar running for a sample roofing
company. [Share it.]

The next step would be a short proposal with two or three options and
pricing. Can we look at it together on [day] at [time], with [other decision
maker] on the call?

Book the follow-up meeting before you hang up. A proposal reviewed together on a call gets decided much faster than one sent by email.

If it isn't a fit

Say so. "Based on what you've told me, I don't think this would pay for itself for you yet. If lead volume picks up, let's talk again." Being honest builds trust, and that owner may refer someone who is a fit.

After the call

• Log the call in your CRM with the numbers they gave you.
• Send a short recap email the same day with the process, the numbers and the next meeting time.
• Write the proposal using their words and their numbers. Our AI automation proposal template shows how.

Practice the full process

The Stride AI Sales Process course covers the outreach, scripts and offer structure used to close AI deals at Executive Stride. Bring a recording or notes from a real call to the weekly live calls for feedback. You can also start with the free guide library.

FAQ

How long should a discovery call be?

About 30 minutes is enough for most small business owners. That's long enough to map the process, put a number on the problem and qualify, and short enough that busy owners will agree to it. Book a separate meeting for the proposal.

Should I show a demo on the discovery call?

Show it at the end, once you understand their process, and connect it to what they told you. Showing it first turns the call into a pitch before you know whether you can help.

What if they ask for the price on the call?

Give the structure and a rough range if you have one, then explain the proposal will have exact options based on what they need. Avoid quoting a firm number before you understand the scope.

What if the decision-maker isn't on the call?

Ask who else is involved and invite them to the proposal meeting. Sending a proposal to someone who has to sell it internally rarely works.