How to research a sales account with AI

Before you begin

Work toward: A concise account brief with sources, hypotheses, and questions.

One target account and a reason to research it
Public company pages and recent, dated announcements
A specific customer problem you can help address

Step 01

Define the question before collecting facts

Write the decision the research should support: whether this account fits, who might own the problem, or what to ask in a first conversation. Set a time limit and name what would disqualify the account. A long company summary is not the same as a useful answer.

Step 02

Build a small evidence file

Collect the company’s own product pages, leadership information, filings where available, and relevant announcements. Record the URL, publication date, and the exact claim each source supports. Use only information you are authorized to share with your chosen AI service. If the service cannot browse, give it the source material rather than asking it to pretend it has searched.

Step 03

Ask AI to organize, not fill the gaps

Have AI group the supplied evidence into company priorities, possible operational changes, and missing information. Require a source beside each factual statement. Keep a separate column for your interpretation. A new job listing might suggest an investment in a function; it does not establish the team’s budget, pain, or buying intent.

Step 04

Turn hypotheses into discovery questions

Choose two or three uncertainties that matter to fit. Convert each into a neutral question a buyer could answer. Instead of asserting that a team has a broken handoff, ask how they measure the handoff today and where delays occur. Draft outreach around one verified observation and one relevant question.

Step 05

Review the brief before using it

Open every source that changes your conclusion. Check the company identity, dates, role titles, and whether an announcement still applies. Remove unsupported personalization. End with a fit hypothesis, the evidence that might overturn it, and the next conversation you want to have. Update the brief after discovery rather than treating the AI draft as a record of fact.

Hypothetical example

What this looks like in practice

The starting point

An AI summary says a company is struggling to scale after announcing a new office.

A more useful approach

The company announced a new office on a recorded date. Capacity problems and purchase intent are unknown. Discovery question: how will the new location change ownership of customer handoffs?

Before you call it done

Every factual claim links to a source you opened.
Inferences and unknowns are visibly separate.
The next step is a question, not an invented customer problem.

Explore all how-to guides

Build a small evidence file

Ask AI to organize, not fill the gaps

Turn hypotheses into discovery questions

Review the brief before using it

What this looks like in practice

The starting point

An AI summary says a company is struggling to scale after announcing a new office.

A more useful approach

The company announced a new office on a recorded date. Capacity problems and purchase intent are unknown. Discovery question: how will the new location change ownership of customer handoffs?

Before you begin

Define the question before collecting facts

Build a small evidence file

Ask AI to organize, not fill the gaps

Turn hypotheses into discovery questions

Review the brief before using it

What this looks like in practice

The starting point

A more useful approach

Before you call it done

Continue reading

Before you begin

Define the question before collecting facts

Build a small evidence file

Ask AI to organize, not fill the gaps

Turn hypotheses into discovery questions

Review the brief before using it

What this looks like in practice

The starting point

A more useful approach

Before you call it done

Continue reading