How to prepare for a first sales meeting when everyone uses AI

Prepare for a first sales meeting the way I wrote about it in 2012: research the company, decide what you want from the meeting, keep your introduction to about 30 seconds and plan the shape of the conversation. What's changed is that AI now does most of the research in minutes, so everyone you're up against turns up prepared, and your edge comes from the parts of preparation a machine can't do.

Getting a first meeting is still hard, and plenty of people still waste it once they're in the room.

The four steps still hold

An old boss drilled the five Ps into me years ago: proper planning prevents poor performance. It took me a good while to admit he was right, which was annoying, because he was smug about most things already. The four steps I wrote down in 2012 were my version of it:

1. Do your research. Know what they do, how they describe their value, and what your CRM says about any history between you.
2. Have an objective. Know what a good outcome looks like before you walk in, even if it's only a second meeting.
3. Keep your introduction short. Who you are, why you're relevant to them and what the meeting is for. Thirty seconds, then it's about them.
4. Plan the structure. Have your opening question ready and know what you want to find out, so your questions fill in the gaps.

I'd still teach all four, and underline the rule inside the first one: never open with "so, what do you do?" In 2012 I called that insulting. With the tools every rep has today, it's unforgivable, because it tells the buyer you couldn't spare ten minutes before asking for their time.

Everyone turns up prepared now

The research I used to spend an afternoon on is close to free. Salesforce's 2026 State of Sales survey of 4,050 sales professionals found 87% of sales organizations already use some form of AI, and sellers expect AI agents to cut prospect research time by 34% once they're fully in place. Your competitors have the same tools, pointed at the same accounts.

Buyers are doing it too. In a Gartner survey of 645 B2B buyers, 45% said they used generative AI in a recent purchase, mostly to gather information on vendors and products. The same survey found 69% prefer to check those AI-generated insights with a sales rep. Gartner also found 51% think they're more likely to run into misleading information from generative AI, and 49% said the same about sales reps, so nobody gets the benefit of the doubt.

So the buyer across the table has often asked a machine about you already, and wants a person to tell them what's true. That's a good position, as long as you bring more than the briefing they generated. (I wrote more about this in buyers still want a human.) When everyone can produce that briefing in seconds, the advantage moves to judgment: knowing which of the ten things it surfaced is worth raising with this person.

What to hand to AI

Give the machine anything that's gathering and summarizing. A good prompt or research agent should get you most of this in minutes:

• The company. What they sell, who to, how they make money, and how they describe themselves on their own site.
• What's changed lately. New offices, acquisitions, leadership changes, product launches, a new system going in. It's the part people skip, and the part that bit me (more below).
• The people you're meeting. Role, time in the job, where they worked before, and anything they've said in public about their priorities.
• Your own history with them. Old opportunities, past contacts and support tickets in your CRM, summarized in a paragraph.
• A draft agenda. A short one you can send ahead, so the buyer knows what the time is for.

I put the warning bluntly in Curiosity: these tools lie with total confidence. AI will hand you a wrong fact about a business as smoothly as the right ones. Check anything you plan to say out loud before you say it, because repeating a confident mistake to a buyer does more damage than saying nothing.

What only you can do

The machine can tell you a lot about a company. It can't tell you what this person will care about on a Tuesday afternoon. That part is yours, and it comes down to three things:

• A hypothesis about their problem. Given their role, their industry and what's changed for them, where is it likely to hurt? Walk in with a sensible guess you're ready to be wrong about, and test it with questions. Know whether you can help at all, and where your own knowledge runs out, so you can say "I don't know, let me find out" without flinching.
• The questions worth asking. Research earns you a better first question. "I saw you've opened a second office and taken on a dozen people this year. That kind of growth usually plays havoc with how work gets tracked. Is that landing on you?" You can only ask that if you did the reading, and it lands very differently from "so, what do you do?" Plan the opening question and the ground you want to cover, and leave the wording loose. Once you're reading lines, you've stopped listening.
• A clear outcome for the meeting. Decide what a good result looks like before you go in. Often it's simply earning the next meeting, with a date in the calendar and the right people invited. Without one, the meeting drifts and ends with "great, we'll be in touch," which means nothing happened.

Then write the short introduction yourself. Mine is some version of "Before I show you anything, I'd love to just understand how things actually work for you at the moment." It tells the buyer this won't be a pitch, which lowers their guard faster than any credentials slide.

The meeting I walked into cold

I learned most of this by getting it wrong. During a brief detour into recruitment, I finally landed a meeting with the CIO of a large insurance company after months of chasing. I knew their tech stack and the roles they hired for. I'd done my homework, just not recently.

What I'd missed was that they'd replaced their entire core software vendor. It had been all over the tech press. I sat down and walked him through candidate profiles built for the old setup, people whose skills were now exactly the wrong ones, and within a few minutes we both knew I was offering answers to a problem he no longer had. A meeting I'd spent months earning was over in about ten.

A change that public is exactly what a five-minute AI briefing picks up, before you've finished your coffee. So refresh the briefing on the morning of the meeting, and read it.

Don't hide behind the briefing

There's an opposite mistake, and AI makes it easier. Research is a respectable place to hide. There's always one more profile to read, one more competitor to check, and now one more briefing to generate. It feels like diligence and can eat a day in which you spoke to nobody.

Prep exists to make the conversation better. Once you have a hypothesis, three or four good questions and a clear outcome, you're ready. If your reps walk in cold, or spend a week reading before every call, that's a coaching problem, and one of the first things I look at as a fractional VP of sales.

I lost that insurance meeting for want of twenty minutes of reading. Nobody gets that excuse now, me included.