Running a Discovery Call: Ask About the Problem, Not Your Product

1. Blog
2. Strategy & Positioning

By Saroj Jha · October 5, 2026 · 9 min read

A good discovery call is mostly questions about the buyer's problem and very little about your product. Put a number on the pain, learn how the decision gets made, and end on a dated next step you both agreed. The five steps below show how, with one example company from start to finish.

A discovery call is the first real sales conversation with a prospect. Its job is to find out whether there is a problem worth solving, whether you can solve it, and whether this buyer can actually buy. It is not a demo. Demo too early and the deal stalls later, because nobody learned what the problem costs or who signs.

Buyers have also changed. In a Gartner survey of 646 B2B buyers run in August and September 2025, 67% said they prefer a rep-free experience, and 45% said they used AI during a recent purchase^[1]. In my view, the call has to give buyers what they cannot get alone: a clearer view of their own problem.

Running case: Example Co. Fictional

Example Co. is a fictional seed-stage B2B SaaS startup whose software automates the month-end close for finance teams at companies with 50 to 500 employees. The controller champions it, the CFO signs, and IT reviews security. It costs about $1,500 a month, its sales cycle is about 45 days, and its team of 8 wants 30 new customers in 12 months. Here, a founder runs a 30-minute first call with Dana, the controller at a 200-person company. All names and numbers are fictional.

Step 1: Open with an agenda and a time check

Start by saying how you would like to use the time, and ask whether that works: "I'd like to spend most of this call on how your close works today, then a few minutes on whether we can help, and the end on a next step, if there is one. Do we still have 30 minutes?" Once the buyer agrees to a plan that is mostly your questions, you have earned the right to ask them. If they now have only 15 minutes, cut straight to the questions that matter most.

Example Co., step 1 Fictional

Dana still has 30 minutes, and she adds one thing: does it work with their ERP (the core accounting system)? The founder writes that down as Dana's question, to answer before the call ends. The founder also asks whether an AI note-taker can join; Dana says yes.

Do this

Write a two-sentence agenda for your next call, with rough minutes for each part, and say it before your first question.

Step 2: Ask about the situation, then the problem, then the impact

Ask in this order, before any product talk. Situation questions learn how things work today (skip what their website tells you). Problem questions find the friction. Impact questions turn a complaint into a reason to spend money.

Research backs leading with questions. A meta-analysis of 155 samples covering more than 31,000 salespeople found that adaptive selling behavior, meaning adjusting your approach to what you learn about each buyer, increased self-rated, manager-rated and objective measures of performance^[3]. You can only adapt to what you have learned by asking. My rule of thumb: if you talk more than the buyer in the first two thirds of the call, you are pitching.

Example Co., step 2 Fictional

"Walk me through your last close." "Where did it slow down?" Dana says reconciliations live in spreadsheets and her two accountants chase other teams for numbers by email. "What did that mean for you and the CFO?" The board pack went out late twice last quarter, and she worked two weekends.

Do this

Write your top two questions for situation, problem and impact, and hold back product talk until all six are answered.

Step 3: Put a number on the pain

A problem without a cost rarely gets budget. For each pain, ask how often it happens, what it costs in hours, money or risk, and what they have already tried. The last tells you what they will not buy again.

Let the buyer do the math, then say their numbers back in one sentence. A neutral reference point helps. APQC, a benchmarking research organization, publishes an open benchmark for the calendar days from the first monthly trial balance to the completed monthly consolidated financial statements; when I checked, the median was 6.0 days across a total sample of 11,223 organizations^[2]. A 12-day close then prompts a fair question: where do the other days go?

Example Co., step 3 Fictional

Dana's numbers (fictional): a 12-day close; she and the two accountants spend about 50 extra hours a month on manual work; loaded cost about $70 an hour. That is about $3,500 a month in staff time alone, against a price of about $1,500. The founder says it back: "Is that roughly right? What would those days be worth to your CFO?"

Do this

For the buyer's top pain, ask "how often, what does it cost, what have you tried", then repeat their numbers back.

Step 4: Surface the decision, and check fit without grilling

Find out how this gets bought: who else is involved, the timeline, and what happens if they do nothing. Do not skip the last one. In a series of decision-making experiments, Samuelson and Zeckhauser found that "individuals disproportionately stick with the status quo," and real choices of health plans and retirement programs showed the same bias^[4]. For a startup, doing nothing is often the strongest competitor.

Weave in the qualification checks, each with one natural question, so the call feels like a conversation, not a form:

What BlackLine's filings show

BlackLine, a public company, sells financial close software, so its annual report shows how this category describes the buyer's problem and its sales motion. Nothing here compares it with the fictional Example Co.

On the problem, BlackLine's annual report for 2024 says: "For organizations of any size, traditional processes are heavily manual and rely upon error-prone spreadsheets, increasing risk and threatening the accuracy of financial reporting"^[5]. On selling, it says "We sell our solutions through our direct sales force," and warns that selling more to existing customers "may require additional sales efforts that are targeted at senior management of such customers, which efforts are often associated with complex customer requirements and additional time to evaluate and test our products, and can lead to long and unpredictable sales cycles." In my reading, that argues for learning about senior people and evaluation steps on the first call. These are what the filings say, not an account of how the company made its choices internally.

If you sell an AI product

In my view, the first call changes in three ways. The buyer will test your claims, so ask what "good enough" means to them: "What would it have to get right before you trusted it?" They will ask what data the tool sees and where it goes, so learn early who must approve that. And they may want proof on their own data before they commit, so ask what a fair test looks like and who judges it.

C3 AI, which sells enterprise AI software, describes this kind of motion in its annual report for the year to April 30, 2026: "Our typical sales cycle begins with one or more product and technical presentations about C3 AI, leading to a mapping of our capabilities to customer use cases." It calls the next stage "initial production deployments, or IPDs": a paid agreement "generally lasting up to six months." It also warns: "Our sales cycles can be long and unpredictable, particularly with respect to large subscriptions"^[6]. This is what the filing says, not how the company decided internally.

Example Co., step 4 Fictional

Dana: the CFO signs anything over $10,000 a year, IT reviews any tool that connects to the ERP, and their outside accountant will ask about audit trails. Doing nothing means another audit season of weekend work. Gaps: no CFO contact yet, IT's steps unknown.

Do this

Write one natural question for each qualification check, and after every call tick off where each was answered.

Step 5: Close on a concrete, agreed next step

End with a specific next step: who, what, when, and what each side brings. "I'll follow up" is how deals stall. Propose it, then ask whether it works. Answer the question the buyer brought, and send a same-day recap: their problem, their numbers, the people and the next step.

Example Co., step 5 Fictional

After answering the ERP question: "Here's what I heard: a 12-day close, about 50 extra hours a month, and a CFO who wants it faster before audit. Could we meet Tuesday at 10 for 45 minutes with you and your CFO? I'll bring a draft close plan built on your numbers, and send our security notes to IT today." Dana agrees and adds the CFO.

Do this

Write your standard closing ask with a day, a time, the people and what you will bring.

Free template

Run this call from one page

The free Discovery Call Guide holds the preparation, agenda, question bank, qualification checks and a note sheet for these five steps, in one Word template.

What changes by stage?

• Pre-seed — Calls are as much learning as selling. Take more notes on the problem than on qualification.
• Seed — Founders still run most calls. Write down the flow and question bank so your first sales hire can run the same call.
• Series A–B — Several people run calls. Make the checks CRM fields and review sample calls for talk time and next steps.

What mistakes should I avoid?

• Turning discovery into a demo in the first ten minutes.
• Accepting a pain with no number on it.
• Meeting only the champion and never learning who signs.
• Ignoring "do nothing" as the main alternative.
• Ending with "I'll follow up".

Example Co. is a fictional company. Example Co., Dana and all their numbers are illustrative.