How I Use AI to Prepare for a Seller Call Before Buying a Small Business

The first call with a business owner is the most underrated moment in a small business acquisition. It is where the seller decides whether you are a serious buyer or another tire kicker, and where you decide whether this business deserves the next 40 hours of your diligence. I use AI to prepare for that call in about 45 minutes, so I walk in knowing the business, the market, and the questions that matter, and I walk out with a clear next step instead of a vague "let's stay in touch."

This post is the exact system. It covers what I research before the call, how I build my question list, what I listen for while the owner talks, and how I turn my notes into a follow-up the same day. AI does the preparation and the organization. I do the listening, the relationship, and every judgment about whether to keep going.

Why the first seller call matters so much

Most buyers treat the first call as an information grab. They show up with a generic list of 30 questions from a blog post, ask about revenue and profit in the first five minutes, and wonder why the owner goes quiet.

Owners of small businesses are not selling a spreadsheet. Many of them spent 15, 20, or 30 years building something with their name on it. A first call that feels like an audit tells them you see the business as a number. A first call that shows you understand their market, their customers, and the work it took to build tells them you might be the right person to hand it to.

That matters for price and terms. Sellers who trust the buyer are more open to seller financing, more patient during diligence, and more forthcoming when something in the numbers looks strange. In my experience, the quality of the first call shows up months later in how smoothly the deal closes.

The first call also protects your time. A good call can tell you within 30 minutes that a business is not for you: the owner is the only person customers will talk to, the revenue depends on one contract, or the seller's expectations are wildly off. Killing a bad deal early is one of the highest value things a buyer can do.

Step 1: Build a one page business brief

Before any call, I build a one page brief on the business. I feed AI everything I have: the listing or teaser, the broker's notes if there are any, the company website, public reviews, social profiles, and any public records I can find. Then I ask for a structured brief with these sections:

• What the business does, in plain language, including who the customers are and how they buy.
• How long it has been operating and any visible signs of growth or decline, such as new locations, review volume over time, or hiring posts.
• Customer sentiment, summarized from reviews, including recurring praise and recurring complaints.
• Owner visibility, meaning how much the owner shows up in reviews, marketing, and the website. If every five star review mentions the owner by name, that is a transition risk worth noting.
• Questions the public information raises, such as a gap in review history, a service listed on the website but missing from the listing, or pricing that looks out of step with competitors.

I read the brief and correct anything wrong. AI is good at organizing scattered information quickly. It is not good at knowing which facts are stale, so I check dates and treat anything unverified as a question, not a conclusion.

This step takes about 15 minutes, and it changes the tone of the call immediately. When I can say "I noticed your reviews mention the same technician by name over and over, tell me about her," the owner knows I did the work.

Step 2: Research the market around the business

Next I ask AI to help me understand the local market. For a service business, that means competitors within the service area, how they price if pricing is public, and how many reviews they have compared to the target. For a business tied to a location, it means traffic patterns, nearby development, and anything that could change demand in the next five years.

I also ask for the obvious risks in the industry right now. Labor shortages, new regulation, a big national player moving into the area, a shift in how customers buy. I do not take these as facts about this specific business. I use them to build better questions.

My prompt is usually something like: "Here is a business brief. List the five biggest external risks to a business like this in this market over the next three to five years, and for each, give me one question to ask the owner that would show whether they have already dealt with it."

That last part matters. The goal is not a research report. The goal is questions that reveal how the owner thinks.

Step 3: Build the question list in the right order

This is where AI saves me the most time. I give it the business brief, the market notes, and my standard question bank, and I ask it to build a call plan in three phases. The order is deliberate.

Phase 1: The owner's story (first 10 to 15 minutes). How they started, what they are proudest of, what a normal week looks like for them, and why they are thinking about selling now. These questions build trust and reveal motivation. The answer to "why now" shapes everything about structure later.

Phase 2: How the business actually runs (next 15 to 20 minutes). Who does what, which customers matter most, how work comes in, what happens when the owner takes a week off. These questions reveal owner dependence, key employee risk, and customer concentration without feeling like an interrogation.

Phase 3: The numbers and the next step (final 10 minutes). Rough revenue and profit trends, what they would want from a buyer, how they think about price and timing, and what they need to see from me to share more detailed financials.

I ask AI to flag any questions from my bank that are risky on a first call. Asking for tax returns before you have built any trust, for example, usually ends the conversation. Those questions go into a "later" list for the second call or the diligence request.

The final plan is usually 12 to 15 questions, not 30. I would rather go deep on fewer questions than rush through a checklist.

Step 4: Know what you are listening for

A question list is only half of the preparation. The other half is knowing what signals to listen for. Before the call, I ask AI to write a short listening guide for this specific business. It typically includes:

• Motivation signals. Retirement, health, burnout, a new opportunity, a partner dispute. Each one points to a different deal structure and timeline.
• Owner dependence signals. Phrases like "customers really only want me," "I handle all the bidding," or "I'm the only one who knows the system."
• Concentration signals. Mentions of one big client, one supplier, or one referral source that drives most of the work.
• Flexibility signals. Openness to staying on for a transition, carrying a note, or keeping a role after the sale.
• Red flags. Vague answers about profit, reluctance to explain a recent dip, or a story that changes when you ask the same question a different way.

I keep this guide open during the call. It helps me notice when an answer deserves a follow-up instead of moving on to the next question. The best information on a seller call almost always comes from the second question, not the first.

Step 5: Run the call like a human, not a script

On the call itself, AI stays in the background. I record only with the owner's permission, and I always ask. Many owners are fine with it. Some are not, and I take notes by hand instead.

I let the owner talk. If they spend 20 minutes on the story of how they started, I let them, because that story usually tells me what they care about protecting. The question plan is a map, not a script. If the call goes somewhere useful, I follow it.

A few rules I hold myself to:

• I never make an offer or name a price on the first call. I do not have enough information, and anything I say becomes an anchor.
• I never promise terms. I can say I am open to creative structures, but I do not commit to anything until I have seen the numbers.
• I always end with a specific next step. Usually that is a request for summary financials, a second call with a key employee, or a site visit. "Let's stay in touch" is not a next step.

Step 6: Turn the call into a structured record

Within an hour of the call, I turn my notes or the transcript into a structured record. I ask AI to sort what I heard into these buckets:

1. Facts the owner stated, with any numbers they mentioned.
2. Motivation and timeline in their own words.
3. Risks I heard, sorted by owner dependence, customer concentration, employees, operations, and market.
4. Open questions I still need answered.
5. Things that conflict with the listing, the broker's notes, or my research.
6. My agreed next step and the date I committed to.

That fifth bucket is the most valuable. When the owner says the business does about 60 percent commercial work and the listing says 80 percent, I want that flagged immediately so I can ask about it calmly later. AI is very good at spotting these inconsistencies across documents, and it does it without the emotional pull of a call that felt great.

I then add my own judgment section, which AI does not write: Do I like this owner? Would customers stay if they left? Is this business worth the next step? Those are my calls to make.

Step 7: Send the follow-up the same day

The follow-up email goes out the same day. I ask AI to draft a short note that thanks the owner, reflects back one or two specific things they said that stood out to me, confirms the next step, and lists exactly what I am asking for.

Then I edit it so it sounds like me. I cut anything that sounds like a template. I make sure the specific details are accurate. A follow-up that reflects back something real, like the owner's pride in a technician who has been with them 12 years, tells them I was listening. A generic follow-up tells them I have a mail merge.

Speed matters here too. A seller who gets a thoughtful note within hours is more likely to send financials quickly. A seller who hears nothing for a week assumes you lost interest.

Step 8: Log it in your deal pipeline

Every seller call goes into my deal pipeline with the structured record, the follow-up, and the next step date. This is where AI earns its keep over months, not minutes. When I talk to the same owner again in three weeks, I can pull up exactly what they said, what I promised, and what is still open.

It also lets me compare businesses on the same terms. After 10 or 20 seller calls, patterns jump out: which industries have the most owner dependence, which brokers send listings with accurate numbers, which motivation types tend to lead to seller financing. That pattern recognition is how a buyer gets sharper with every deal, whether or not that deal closes.

If you want to see how I think about the full acquisition workflow, from sourcing through diligence, my operator stack and AI implementation approach cover the systems behind it. And if you are earlier in the process, my post on using AI to source and screen businesses to buy is the right starting point.

Common mistakes buyers make on seller calls

After many owner conversations, these are the mistakes I see most often, including ones I made early on.

Leading with numbers. Asking for revenue and profit before the owner trusts you tells them you are a spreadsheet buyer. Get the story first.

Talking too much. Buyers who spend the call explaining their background, their fund, or their vision rarely learn anything. The owner should talk at least 70 percent of the time.

Treating the broker's summary as fact. Brokers summarize for marketing. Use the call to verify, gently.

Forgetting the owner's real goal. A retiring owner who cares about their employees will weigh that over a few percent of price. If you never ask what they want for their team, you will never know.

Letting AI talk for you. AI drafts my follow-ups and organizes my notes. It never decides what I say to an owner or whether a deal is worth pursuing. Owners can tell when a message was written by a machine, and in a relationship business, that costs you.

No next step. Every call should end with something specific scheduled or requested. Without it, deals drift and die.

What this system costs and saves

The tools behind this are simple. A general purpose AI assistant, a transcription tool when the owner consents, and a pipeline tracker. For most buyers, that is well under $100 a month.

The savings are in time and quality. Preparation used to take me two to three hours per seller and still felt scattered. Now it takes about 45 minutes, and I walk in with better questions. The structured record and follow-up take 15 minutes instead of an evening. Over a search where you might talk to 30 or 40 owners, that is the difference between a buyer who burns out and one who keeps showing up sharp.

The bigger saving is in deals I do not chase. When the structured record shows customer concentration over 40 percent and an owner who personally holds every key relationship, I can pass early and politely, instead of spending weeks in diligence to learn the same thing.

Prepared call vs unprepared call: side by side

The post-call scorecard I fill in

After every call I score the business on a simple five-line scorecard before I decide whether to keep going. Each line gets green, yellow, or red.

1. Owner dependence: Does the business run when the owner is away, or does every key relationship sit with them?
2. Customer concentration: Is any single customer a large share of revenue? Over 40 percent is a red line for me.
3. Records quality: Are customers, jobs, and invoices in one system, or spread across notebooks and phones?
4. Seller motivation: Is the reason for selling clear and consistent across the conversation?
5. Fit with my systems: How much of the weekly work is repeatable motion I know how to systematize?

Two reds and I pass politely. Mostly green and the deal gets a folder and a second call.

A sample seller call prep packet

Here is what the finished packet looks like for a typical call. Everything fits on two pages:

• Page 1: The business brief, the top five market risks, and the five things I most want to learn.
• Page 2: The 12 to 15 question plan in three phases, the listening guide, and the specific next step I want to propose at the end.

After the call, the packet gets a third page: the structured record, my judgment notes, and the follow-up I sent. That three page file becomes the start of the deal folder if the business moves forward.

FAQ

What should I ask on the first call with a business owner?

Start with the owner's story and their reason for selling, then ask how the business runs day to day, who the key customers and employees are, and what happens when the owner is away. Save detailed financial requests for the end or for a second call once trust is established.

Should I record a seller call?

Only with the owner's clear permission, and follow the recording consent rules where you and the owner are located. Many owners agree. If they decline, take notes by hand and build the structured record right after the call.

Can AI tell me whether a business is a good deal?

No. AI can organize information, spot inconsistencies, and help you prepare better questions. Whether a business is worth buying depends on judgment about the owner, the customers, the risks, and your own goals, and that decision stays with you.

How long should a first seller call be?

Plan for 45 to 60 minutes. That is enough time for the owner's story, a real conversation about how the business runs, and a clear next step, without exhausting either of you.

When should I talk about price on a seller call?

Not on the first call. Ask what the owner hopes to achieve and what a good outcome looks like for them, but hold any discussion of price or terms until you have seen summary financials and understand the business.

What is the most common red flag on a first seller call?

Vague or shifting answers about profit, especially when combined with heavy owner dependence. If the owner cannot explain how the business makes money without them, the transition risk is high.

Build the system once, use it on every deal

Buying a small business is a relationship business that happens to involve a lot of documents. AI handles the documents, the research, and the organization so you can show up fully present for the part only you can do: listening to an owner describe what they built and deciding whether you are the right person to carry it forward.

If you want help building an acquisition operating system like this for your own search, you can see how I work with operators on AI implementation.

Current Search Intent Check

Recent Search Console data shows people arriving through "automated real estate investing systems". That changes the bar for this post: it needs to answer the operator question directly, name the workflow being improved, and give the reader a practical decision rule instead of another broad AI opinion.