The Current

Opportunity intelligence newsletter

Issue #002

Signal 03 · AI Visibility

The AI Mention AuditOne week. Then a verdict.

One buyer. One offer. A test that is allowed to fail.

Issue #002 · October 2026 · 8 min read

The brief

Issue #001 was the map. This one runs a road.

We picked AI visibility because one person can test it without a client, a product, or a team. Implementation and agents may be larger. They are worse first tests. If you need someone else’s company before you learn whether the idea is real, you are pitching, not testing.

The reader is the same as Issue #001: a creator or operator who sells a service and gets work by being found. Not an SEO department. Not a lab.

The business people will want to start is an agency. We are going to see if a much smaller offer deserves a week.

  1. Signal

    The short list is the new front door.

    When someone asks an assistant who to hire, the answer is a handful of names, not ten links. If you are not named, you are not in the argument. That is the change. It is not proof that search is dead. It is a second door, and almost nobody has opened theirs.

  2. Problem

    Owners don't know what the model says.

    A creator or operator who sells a service has no report for this. Search consoles don't answer it. The offers already in market are monthly “GEO” retainers that look like SEO with the label swapped. The pain is smaller than a retainer: what happens when a stranger asks?

  3. Opportunity

    Sell the answer. Not the repair.

    Buyer: an owner-operator who already gets work by being found. A studio, a consultant, a local firm, a creator who sells done-for-you work. They have a public email. Offer: eight prompts a real buyer would type, run in ChatGPT, Gemini, and Perplexity. One page back — named or not, what was said, three changes inside their control. Price for the test: $250. No retainer in the same email. $250 is expensive enough that a yes means something, and cheap enough that a stranger can decide without a call.

The score

Five factors. A 1 kills it.

Each factor is 1 to 5. Five is favorable. Our bar, set before the points: under 15, stop. 15 to 19, run the week and build nothing. 20 or higher, run the week and design the second offer.

  1. 3

    Buyer urgency

    They don't feel it until they see the screenshot. The email has to create the urgency.

  2. 3

    Willingness to pay

    $250 is a test price. A monthly service is unproven, so it is not part of the test.

  3. 5

    Speed to a test

    One person. One week. No product, no partner, no one's permission.

  4. 2

    Crowding

    The retainer version is already a pitch. The narrow audit is less crowded, and easy to copy.

  5. 4

    Can you deliver it

    The audit is an afternoon. The fix is not. Do not sell the fix in the test.

3 + 3 + 5 + 2 + 4 = 17. No factor is a 1. The offer stays alive. The company does not.

The experiment

Seven days. Fifteen emails. Then stop.

  1. Day 1–2

    Pick one category you know.

    One place, or one specialty. “Wedding photographer in Brooklyn.” “Newsletter strategist.” Not “AI.” Write eight prompts a buyer would actually type, including the constraint and “who should I hire.”

  2. Day 3

    Run the prompts. Don't argue.

    ChatGPT, Gemini, Perplexity. Save the answers. Mark who got named. You are collecting a fact, not a strategy.

  3. Day 4

    List 15 people who were absent.

    Businesses or creators you can email without a warm introduction. Public address only. If you can't find 15, the category is wrong. Stop.

  4. Day 5–6

    Send one email.

    The prompt. The fact they were not named. One screenshot. No deck. Ask if they want the full page for $250. Then stop writing.

  5. Day 7

    Count. Don't revise.

    Pass line, written before you send: 3 replies from 15, or 1 person who pays. Anything less and this offer is a no. The signal can still be real. The offer is what failed.

Validation

Test. Don’t build.

17 / 25

Crowding is a 2 because the retainer version of this is already a slide in every marketing Slack. The audit is narrower. That is the only reason it is still alive.

What would change the verdict: two paid audits from people you did not already know, and one of them asking what it costs to change the outcome. Until then, do not name an agency, do not spec a tool, and do not write another visibility issue.

This does not prove three things:

  • That assistants have replaced search for your category.
  • That being named produces revenue.
  • That the people who reply are the people with a budget.

If you run the week, the only number that matters is sent, replied, paid.

A hypothesis from The Current. Not a business plan, and not a recommendation to quit anything.

Sign up

Get the next issue in your inbox.

The Current is an opportunity intelligence newsletter for creators and operators. One issue when a signal is worth running. Hosted for beehiiv delivery.

One issue when it matters. Built for beehiiv delivery. Unsubscribe anytime.

Also

The map is Issue #001.

Issue #001 names the shift and the method. This issue is the first time the method is allowed to fail.

Read Issue #001

Prefer a dedicated signup page?

Open the subscribe page