How to Research a Prospect Before You Call

The ten minutes that change the call

There are two versions of you on a call. One skimmed the company's about page in the lift on the way up. The other knows they opened a second warehouse in March, that their finance director started six weeks ago, and that everyone in their industry is currently worried about the same piece of regulation.

The second version did not work harder. That version spent ten minutes with an AI assistant and a browser tab, and came out with better questions.

Research is where AI is genuinely excellent, and also where it will mislead you with a completely straight face. Both things are true at once, and this lesson is about holding them together.

The five things worth knowing

Before a first conversation you want five things. Not twenty. Five.

What the company actually does, in plain words you could repeat back to them. What has changed recently: funding, a new office, a launch, a merger, a rough quarter. What pressure their industry is under right now. Who you are meeting and what their job makes them care about. And your angle, which is the sentence connecting all of that to what you sell.

Ask for all five in one go.

I have a first call with [name], who is [job title] at [company].
Company website: [url]. We sell [one line about your product].
Give me a brief with five sections, short bullets only:

What this company does, in plain language
Anything notable that has changed in the last 12 months
Two or three pressures their industry is facing right now
What someone in this job title typically worries about
Three questions I could ask that show I have done my homework
For every factual claim about this specific company, mark it CONFIRMED
if you can tell me where it came from, or UNVERIFIED if you cannot.

That last instruction is the whole lesson in one line. Make the tool show its working and the gap between what it knows and what it is improvising becomes visible.

Verification is the job

Here is the rule: anything specific enough to be impressive is specific enough to be embarrassing. Numbers, names, dates, launches, headcount. Those are the ones to check.

Checking is faster than you expect. Open the company's own site, their news or press page, and their leadership page. If ChatGPT told you their operations director is called Sam and the leadership page says Priya, you have just saved yourself a genuinely awful opening minute. If your tool cannot browse the web at all, treat every company-specific claim as a guess dressed as a fact, because that is exactly what it is.

Turning research into questions

The common mistake is using research to talk. The point of research is to ask.

Nobody is impressed that you read their annual report. They are impressed when your question shows you understood what it meant for them.

Here are the confirmed facts I have about a company I am meeting.
FACTS:
[paste only the ones you checked]
Give me six open questions I could ask on the call. Each one should be
answerable only by someone who works there, and should show I understood
the implication of a fact rather than reading it back.
No questions that could be asked of any company. No yes or no questions.

When you know the person, not just the company

If they have given a talk or written something public, feed it in. If a colleague has notes from a previous call, use those instead, with customer names and account numbers stripped out first.

Here is public background on the person I am meeting: [paste the bio,
the talk summary, or the article they wrote].
Tell me what they seem to care about professionally, what language they
use for their own work, and one thing I should avoid assuming about them.
Keep it to eight lines. Flag anything you are guessing at.

Then close the notebook. You are not going to recite any of this. You are going to sound like someone who paid attention, which is rarer and much better.

There is a line here worth stating plainly, because the rest of this course keeps returning to it. Using AI to prepare better is good and always has been. Using it to fake personal knowledge you do not have, name-dropping a mutual contact you have never met, referencing a report you did not read, is a different thing entirely. It damages your reputation the first time you get caught, and buyers catch it more often than sellers think.