How consultants prep for a discovery call in 20 minutes using AI

The call that decides whether an engagement happens usually gets the least-prepared hour of the week: a website skim, six LinkedIn tabs, and a question list that could be asked of anyone. This guide replaces that with a research workflow that digs into the prospect’s company, the person you’re meeting, and their industry’s this-quarter pressures — then builds a question set specific enough that none of it could be asked of anyone else.

After a one-time setup, your prep for any call is one message:

“Prep me for Thursday’s call with [name] at [company]. They reached out about repositioning — here’s their email.”

Why this isn’t just a ChatGPT prompt

  • It pulls live data a plain chatbot can’t reach. LinkedIn company pages, personal profiles, grounded web search, community sentiment — real lookups with sources, not training-data guesses about a company that changed last quarter.

  • It knows your practice, so the brief includes fit. A one-time business-profile skill means every brief answers “where does this prospect fit what I sell” — the section generic research can’t produce.

  • It recovers from dead ends itself. Wrong LinkedIn slug? Private profile? It searches for the correct page and retries. You give it a name, never URLs.

  • It flags bad data instead of repeating it. In our documented run it caught a LinkedIn artifact claiming 593 employees at a 10-person company — and put it in the watch-out section rather than the brief. A generic prompt would have quoted it confidently.

  • Questions come organized by purpose, not as a flat list. Openers, problem-depth probes, industry-specific questions, stakeholder mapping, closers — a strategy you pick from, not a checklist you recite.

The steps

  1. 1.One-time setup: teach the AI your practice (~30 min)Pro

    Run the Business Profile Extraction Prompt. It interviews you about your positioning, offer, and typical engagements, and saves the result as a skill every future prep loads automatically.

    Say this

    “Find the Business Profile Extraction Prompt on Amplifiers and run it.”

  2. 2.When a call books: start the workflowPro

    Run the Discovery Call Prep Workflow. It asks which research you want (company, person, web, Reddit, competitor ads) and what the situation is — inbound, outbound, cold, or networking. That distinction changes the entire question set it builds.

    Say this

    “Run the Discovery Call Prep Workflow for my call with [name] at [company]. They reached out — here’s their email.”

  3. 3.Let it dig (you do nothing here)

    It runs its own searches: the company in its own words, the person and what they’re publicly engaging with, industry pressures, news. When a lookup misses, it finds the right page and retries; when something genuinely can’t be found, it says so in the brief.

  4. 4.Read the brief as the expert

    Company, person, industry, fit read, and a watch-out list. You know your market better than any search — kill what smells off and ask follow-ups on anything thin. Your corrections sharpen the questions.

  5. 5.Take the questions, not a script

    Pick the eight you’d actually ask in the time you have. The call stays yours — the difference is you’re listening instead of scrambling.

What it looked like when we ran it

We ran the full chain on a real prospect company (anonymized) and published the brief. Three moments from that run:

It found the opening question no website shows

The founder had publicly engaged with a user describing finding the product through an AI assistant — and the signup flow asking which prompt brought them there. Cross-referenced with industry research on AI-driven discovery, that became the first question of the call.

Two lookups failed; the workflow fixed both itself

A wrong company slug and a private profile URL. It searched for the correct pages and retried without asking for links — and the guide keeps that in, because the recovery behavior is the product.

The watch-out list earned its place

LinkedIn claimed 593 people “associated with” the company; the real team is ten. Flagged as a data artifact — because citing a wrong number on a call is worse than no number.

The consultant in the sample is a composite we invented; the prospect company is real (anonymized) and every finding, quote, and number is genuine tool output.

Common questions

How is this different from asking ChatGPT to “research a company”?

Reach and honesty. The workflow drives live LinkedIn, web, and community lookups with sources — and when a lookup fails or data looks wrong, it flags it instead of hallucinating around it. It also cross-references everything against your practice, which generic chat can’t do.

What if my prospect has almost no public footprint?

You get a thinner brief that says exactly which sources came up empty — useful signal in itself. It reports gaps transparently rather than padding.

Do I need to collect LinkedIn URLs first?

No. A name and a company are enough — it finds the pages itself and retries when its first guess is wrong.

How long does prep actually take?

In our documented run: about twenty minutes, most of it unattended. Your part is reading the brief and picking your questions.

Run this guide before your next discovery call

The Discovery Call Prep Workflow and Business Profile Extraction Prompt are in Amplifiers Pro.

Try Amplifiers Pro today

Works inside Claude · setup takes minutes · cancel anytime