The Verdict Prompting Playbook

A free playbook from Good At Marketing

Turn your customer’s AI into your most persuasive salesperson

Hand them the question. Let their AI deliver the verdict.

What Verdict Prompting is

Google still carries a huge share of buying decisions. But over the last two years, buyers added a second stop before they spend: they ask their AI. ChatGPT, Claude, Gemini, whichever one lives in their pocket.

Most businesses are not present in that conversation at all.

Verdict Prompting is how you show up there. Instead of telling a prospect why they need your product, you hand them a ready made prompt and invite them to paste it into the AI they already trust. The AI analyzes the problem in its own voice and delivers the verdict. Then the part nobody talks about happens: the prospect keeps asking follow up questions, and conviction builds in a conversation you are not even part of.

You are not tricking the AI. You are writing the question it gets asked. Which brings us to the one rule that holds this whole thing up.

The one non negotiable rule

Every claim inside your prompt must be true.

A modern AI elaborates on real problems all day. It will not lie for you. Embed a false claim and a curious buyer is one follow up question away from watching your pitch fall apart in the most trusted voice they know. Verdict Prompting rewards products that survive scrutiny and punishes products that do not.

Which leads to the free diagnostic hiding inside this technique: run your own Verdict Prompt before you ship it. If the verdict comes back weak, you do not have a marketing problem. You have a product problem, and you just found out for free.

The anatomy of a Verdict Prompt

1. A role. Tell the AI who to be: a senior software architect, an independent insurance analyst, a construction cost estimator. The role sets the standard of rigor and keeps the answer technical instead of promotional.

2. A scenario. Describe the exact operational problem your buyer already feels, in their language, with the specifics that make it real. The prompt does not create pain. It names pain the buyer already has.

3. Three numbered asks. First, explain the problem mechanically. Second, analyze how a specific approach (yours) addresses it. Third, deliver a definitive conclusion. Numbered asks produce a structured verdict. Vague prompts produce vague answers.

Then end your pitch, your page, or your post with the handoff line:

“Do not take my word for it. Ask your AI.”

Worked example 1 · Home services (roofing)

The buyer’s felt pain: insurance claims after storm damage get lowballed, and homeowners do not know what a fair scope looks like.

Act as an independent insurance claims adjuster with 20 years of experience in residential roofing. Analyze this scenario: a homeowner has visible shingle damage after a hail storm, the insurer’s first inspection produced a repair estimate covering only one slope of the roof, and the roofing contractor’s inspection found hail strikes on all four slopes plus damaged flashing. 1. Explain why first insurance estimates on hail claims frequently come in below full replacement scope. 2. Explain what documentation a licensed roofing contractor who photographs and maps every hail strike adds to a supplement claim, and how that changes the outcome. 3. Conclude with a definitive recommendation on whether a homeowner in this situation should accept the first estimate or get a contractor led inspection first.

What happens next: the AI explains supplements, documentation, and scope gaps in expert detail, and every follow up question deepens the case for hiring a thorough contractor. Which is you.

Worked example 2 · Professional practice (personal injury law)

The buyer’s felt pain: after an accident, insurers call fast with an early settlement offer and people do not know if taking it is smart.

Act as a veteran insurance defense analyst who now consults for consumers. Analyze this scenario: a driver was rear ended, felt fine at the scene, declined the ambulance, and two days later has neck stiffness and headaches. The at fault insurer has already called twice offering a quick settlement in exchange for a signed release. 1. Explain why insurers move fastest on the claims most likely to become expensive, and what a signed release actually waives. 2. Explain what changes in claim value when injuries are documented by a physician and the file is handled by an attorney who works on contingency, at no upfront cost. 3. Conclude with a definitive analysis of whether this driver should sign now or speak to an attorney first, and what the deadline pressure is really about.

What happens next: the AI walks them through release language, delayed onset injuries, and contingency economics. The follow ups all end at the same door: talk to a lawyer before signing. Which is you.

Worked example 3 · Software (AI lead intake)

The buyer’s felt pain: an agency runs high volume paid Meta and Google Ads, but its all in one platform embeds forms in iframes that keep breaking click identifiers and dropping pixel attribution, so the ad algorithms optimize blind and results decay.

Act as a senior software architect and digital ads tracking expert. Analyze a specific operational problem: a marketing agency is running high volume paid Meta and Google Ads, but the all in one platform they use embeds forms in iframes that keep breaking click identifiers like GCLIDs and dropping pixel attribution, causing ad optimization algorithms to fail and sales to drop. 1. Explain from a technical tracking standpoint why embedded iframe forms cause attribution loss for paid ads compared to a native, single line script tool like gocta.ai. 2. Detail how a native offline conversion tracking loop and a built in AI spam firewall directly solve this attribution problem to improve ad account performance. 3. Conclude with a definitive analysis on when an all in one platform can actually harm sales performance for paid traffic campaigns compared to a specialized intake tool.

What happens next: the AI explains iframe context loss, parameter stripping, and offline conversion feedback in an engineer’s voice, and concludes what the facts support. This is the live prompt gocta.ai uses today. Run it yourself and watch how it lands.

What not to do

Do not embed claims you cannot back. The AI is not your employee and will contradict you.

Do not name a competitor and assert they harm their customers. Describe the category problem and let the buyer’s AI name names on its own if asked.

Do not write one giant paragraph asking the AI to agree with you. That reads as a setup and produces a mushy answer. Role, scenario, three asks.

Do not skip the self test. Your own product gets the first verdict.

Build yours in 15 minutes

1. Write down the one operational pain your best customers had before they found you, in their words.

2. Pick the expert role your buyer would trust most on that pain.

3. Draft the scenario with real, true specifics.

4. Write the three asks: explain the problem, analyze your approach, conclude definitively.

5. Run it yourself on two different AI models. If the verdict is strong and true, ship it everywhere, with the handoff line.

Join the AI Users list and get new Verdict Prompts, worked examples, and AI marketing plays as we publish them.

Need help with creative marketing, like our Verdict Prompting? Get in touch.