Agency Client Discovery Calls: Prep with AI Panels in 30 Minutes

The discovery call is where most agency engagements get their tone for the next 90 days. Done well, you walk out with a sharp brief, an aligned client, and a price point that respects the strategic work ahead. Done badly, you walk out having spent an hour extracting basic customer information that the client already had in 6 internal documents, and now your team has to spend a week catching up before you can think.

Most agencies do discovery badly. Not because the questions are wrong, but because the agency walks in cold. They have read the client's website and maybe scanned the pitch deck the client sent. They ask discovery questions the client has answered 30 times to 30 different vendors, and the client responds with the same canned answers each time.

In 2026, the sharp agencies do something different. They run a 30-minute synthetic panel on the client's customer before the discovery call. They walk in with 8 specific hypotheses about the customer that they want to validate or correct. The call becomes 2x faster, 3x sharper, and the client ends it thinking the agency already understands their business.

Here is the workflow.

What discovery should actually accomplish

Strip the meeting down to its 4 real goals:

1. Validate or correct your model of the client's customer. What you already believe vs what is actually true.
2. Surface the constraints the client has not put in writing. Budget ceiling, internal politics, regulatory limits, the partner who must approve everything.
3. Align on the actual problem being solved. Often different from the problem in the RFP.
4. Establish your strategic credibility. The client should leave the call feeling like they hired a thinking partner, not a vendor.

Notice that exactly zero of these goals require the client to spend 45 minutes explaining who their customer is. That part should be 90 percent done before the call starts, which is what synthetic panels enable.

The 30-minute pre-call workflow

This is your discovery-prep loop. It runs the morning of the call.

Step 1: Build the customer panel from public signals (10 minutes)

Use everything publicly available about the client to model their customer:

• The client's website (especially the case study and customer logo pages).
• The client's LinkedIn (employee count, fundraising stage, geography).
• 3 to 5 customer reviews from G2, Capterra, or similar.
• Any press coverage that profiles the client's customers.
• The original RFP or brief if you have it.

Feed this material into Custom Audience Builder and spin up a 30 to 50 persona panel that models the client's customer base. If the client serves two distinct segments (say, marketers and sales leaders), build two panels.

Step 2: Generate 8 strategic hypotheses (5 minutes)

Open a doc and write 8 hypotheses about the client's customer that, if true, would change how you would approach the engagement. Vary them:

1. "Their customer's primary decision criterion is X, not Y."
2. "Their customer probably evaluates 3 to 5 competitors and the top alternative is Z."
3. "Price sensitivity peaks around the $X threshold, above which deals slow significantly."
4. "Their customer uses tools A and B daily but not C, despite C being featured in their integration list."
5. "The biggest objection their sales team hears is probably X."
6. "Their primary growth lever right now is X channel, but the data suggests Y would yield higher ROI."
7. "Their customer's job title is officially X but the actual decision- maker is more likely Y."
8. "The thing their customer most wants to be true about themselves is X, which is why they buy."

Make them specific and falsifiable. Hypotheses that the client cannot disagree with are not useful.

Step 3: Test each hypothesis with the panel (10 minutes)

Run each hypothesis through the panel. Ask:

• For each hypothesis, do you agree or disagree? Why?
• If you disagree, what would the more accurate statement look like?
• Which hypotheses, if true, would most change how you choose tools or agencies in this space?

The panel returns confirmed hypotheses, corrected hypotheses, and a ranking of which ones are most decision-influencing. You now have a sharp view of the customer that took 30 minutes to develop and would have taken 2 weeks of customer interviews to approximate.

Step 4: Format for the call (5 minutes)

Write up 3 to 5 of the most interesting findings in 1 to 2 sentences each. Format each as: "Hypothesis we tested + what the panel suggested + question for the client."

For example: "We modeled your customer and the panel suggests that decision authority sits with the VP of Operations, not the Director of Product as your job-title targeting suggests. Is that consistent with your actual sales conversations?"

This format positions you as bringing thought to the call, not extracting information. The client now has something to react to, which produces much sharper input than open questions.

What this changes in the call itself

Three dynamics shift.

First, the meeting is 30 to 40 percent shorter. You skip the entire "explain your customer" section because you have already done the work. A 60-minute discovery becomes a 35-minute validation session, and the client appreciates getting time back.

Second, the client's feedback gets sharper. Instead of generic answers to generic questions, the client is reacting to specific hypotheses. They think harder, they correct you when you are wrong, and they add the detail you could not have known. You learn 3x more per minute.

Third, your strategic credibility lands in the first 5 minutes. The client immediately understands that this is not their usual agency call. You move from "vendor pitching services" to "advisor with a perspective" in the opening, and the entire commercial dynamic of the engagement shifts.

A real example from a brand agency engagement

A brand agency we work with prepped for a discovery call with a B2B SaaS client in the data observability space. Standard pre-reads, standard preparation. Then they added a 30-minute panel test.

The panel surfaced 3 things that changed the meeting:

• The client's stated ICP was "data engineers." The panel suggested the actual buyer was the VP of Engineering, with data engineers as influencers. Their entire site copy was written for the wrong persona.
• The biggest objection their sales team would be hearing was likely about cost-of-ownership, not feature breadth. Their pricing page focused on features.
• Their top alternative in buyer minds was an open-source tool the client did not list as a competitor, which meant their differentiation messaging was missing the actual decision context.

The agency walked in and led with these 3 hypotheses. The client confirmed all 3 within 10 minutes. The remaining 50 minutes became a strategic conversation about how to reposition the brand, instead of an information- extraction session.

The agency closed a $180k engagement on the spot, with a brief written collaboratively in the call itself. Without the panel prep, the same call would have ended with "send us a proposal, we'll review next week" and a 50 percent close rate.

Where this scales for agency workflow

Three patterns to embed in your agency operations:

1. Make panel-prep the default for every discovery call. Build it into your pre-call workflow the same way you build in CRM logging. Junior AEs can run the panel; senior strategists review the output and formulate the hypotheses.
2. Save panels per client. Every panel you build is reusable across the engagement. The discovery-call panel becomes the brief panel becomes the campaign-test panel becomes the QBR insights panel. One investment, used across the engagement.
3. Productize the output. Some agencies are now selling the panel summary as a paid "pre-engagement audit" at $5k to $15k. It is high-margin, fast to deliver, and creates a paid first touchpoint that often expands into the larger engagement.

The honest comparison to real customer research

Synthetic panels do not replace actual customer interviews. They make them unnecessary at the discovery stage and more productive at the brief stage.

The right ratio:

• Discovery prep: 100 percent synthetic panel. You have no time and no budget for real interviews this early.
• Brief development: Synthetic panel for hypothesis generation, then 3 to 5 real customer interviews to validate the most strategic hypotheses.
• Strategy testing: Mostly synthetic panel for breadth, with 2 or 3 customer interviews to add qualitative depth.

The agencies that win in 2026 are not the ones that abandon customer research. They are the ones that use synthetic panels to do the breadth work at a fraction of the cost, so they can invest the saved budget in the customer research that matters most.

What to ship into your agency this week

Pick your next 3 discovery calls. Run the 30-minute panel workflow before each one. After 3 calls you will know whether this changes your close rate, your average deal size, or your client perception. Most agencies that adopt this find all 3 metrics shift in their favor within a quarter.

The discovery call is the most expensive 60 minutes of an agency's week, because what happens in that hour shapes the next 90 days of the engagement. Prepping with a synthetic panel turns it from an information session into a strategy session. That is a structural change in how agency work gets done, and the agencies that adopt it first will outcompete the ones that do not for the next 2 to 3 years.