Step-by-Step Playbook

How We Get Customers From ChatGPT and Gemini Every Month

The $0 Playbook

BT
Ben Tannenbaum
12 min read

Everyone talks about AI search. Nobody explains what to actually do about it. Here's the exact repeatable playbook we use — step by step. It's free, it's simple, and it works.

We consistently increased our mentions in ChatGPT and Gemini by 20%+ month over month, above the general growth curve in AI usage. Here is the exact playbook we use. It's repeatable, and right now it costs nothing.

Step 1: Find Out What People Actually Ask AI About Your Category

Most brands have no idea what real people are typing into ChatGPT or Gemini about their space. That's where everything starts.

I'm a founder at AISO ( getaiso.com), so I'll walk through this with our own tool. Head to the AISO search bar and start with a broad query in your category. For us, that's something like: "getting leads from chatgpt ai search optimization."

AISO search bar showing broad category search
Start with a broad category search to pull thousands of real AI conversations.

This returns a broad sample of thousands of real, anonymized conversations people have had with ChatGPT and Gemini. Real questions, real follow-ups, real answers.

Step 2: Refine With Smart Filters

A broad search gives you volume. Now you need precision. AISO's AI Filter lets you describe, in natural language, exactly what you're looking for.

For example, I wrote: "I am a founder at getaiso.com which helps brands be visible in AI search like ChatGPT and Gemini but focusing on providing samples of real AI searches. I want to know what my potential clients, brands and SEO agencies, are asking AI (ChatGPT, Gemini, etc.) so I can create content to meet those intents."

AISO AI Filter interface
The AI Filter lets you describe your ideal audience and intent in plain language, then filters thousands of conversations down to the ones that matter.

The filter runs across 2,000 results and surfaces conversations that match your described intent. You go from noise to signal in one click.

Step 3: Extract the Gold

After filtering, I get roughly 100 high-intent, content-rich conversations. These are real prompts from real people, like:

  • "How to do SEO optimization for intelligent AI assistants?"
  • "What are the strategies used for search in AI?"

And here's the key: you don't just see the prompt. You see the full conversation. The AI's answer, the user's follow-up questions, and the depth of the exchange.

Real AI conversation example
A real AI conversation showing the user's prompt, the assistant's full answer, and follow-up questions. This is the kind of intent data you can't get from traditional keyword tools.

This is research gold. You now know exactly what your potential customers want to learn, what language they use, and what follow-up questions they have. No keyword tool built for traditional search gives you this.

Step 4: Save, Download, and Cluster Your Prompts

Select the best prompts and save them to your account. You can also download them as a CSV so you keep them forever.

Saved prompts interface in AISO
Your saved prompts in AISO, ready to download or organize.

If you have a lot of prompts, group them into topic clusters. I load the CSV into Claude and ask it to cluster the prompts by theme. A tip: keep some raw prompt examples alongside the clusters. You don't want to over-abstract and lose the real language your audience uses. The signal-to-noise ratio matters.

Step 5: Fill the Funnel Gaps

The prompts you find in AI conversations tend to be top-of-funnel and mid-funnel. That's typical for AI search. But you need the full picture.

For bottom-of-funnel questions, think about what your sales team hears over and over: "Where do you source the data?" "Is it GDPR compliant?" "How does pricing work?" Sales call recordings are great for this.

You can also check People Also Ask boxes and Reddit, but remember: people ask very different things to ChatGPT than they do publicly. They're more honest, more specific, and more exploratory with AI. We often say people share more with ChatGPT than with their romantic partners.

Step 6: Discover What AI Actually Searches (Query Fan-Out)

This is the part nobody talks about.

When you ask ChatGPT a question, it doesn't just "know" things. It runs multiple background searches to assemble its answer. We call this Query Fan-Out, and understanding it changes everything about your strategy.

In AISO, go to the Fan-Out feature and generate the fan-out for your saved prompts. You'll see the exact queries ChatGPT and Gemini run behind the scenes:

  • "How to do SEO optimization for intelligent AI assistants"
  • "SEO for AI assistants optimization tips AI search optimization"
  • "Search strategies used for search in AI"
  • "In AI search, uninformed search strategies list what are they"
Fan-Out Details table showing background queries
The Fan-Out Details table shows each background query, its priority level (P1, P2, P3), and whether your site or competitors rank for it.

Notice the priority column. P1 queries are the ones ChatGPT searches first. That's where your content has the highest leverage.

Notice also the competitor columns. The app tells you who currently ranks for each fan-out query. In this example, neither our site, Peec.ai, nor Tryprofound are ranking for these queries. That's a wide-open opportunity.

Step 7: See What's Actually Ranking (And Where You Need to Be)

Click "Show" on any fan-out query to see the actual Google and Bing results for it. This is where the strategy becomes concrete.

Fan-out query search results
Expanding a fan-out query reveals the actual search results that ChatGPT pulls from. These are the pages you need to outrank.

These are the pages ChatGPT is pulling from to build its answers. If you want to be mentioned, you need to be here. The strategy: create content that ranks for these specific queries on both Google and Bing.

This is a relatively easy and affordable strategy to get into ChatGPT's answers. There is a lot more to it, but as a DIY starting point, this is powerful.

Step 8: Create Content Hitting Those Exact Intents

Now you create content that targets the exact queries AI runs when answering prompts in your category.

Not guessing. Not using keyword tools built for 2019 SEO. Using the actual queries that ChatGPT and Gemini search for, derived from real conversations in your space.

You already have:

  • The real prompts people ask (from Steps 1-5)
  • The exact queries AI runs to answer them (from Steps 6-7)
  • The pages currently ranking for those queries (from Step 7)

Write content that answers those queries better than what currently exists. Optimize for both Google and Bing. Publish, build links, and let the compounding begin.

Step 9: Track Your Success

Start simple. In Google Analytics, look at sessions by source and filter for referrals from chatgpt.com and gemini.google.com. You should see a trend like this:

Google Analytics showing ChatGPT and Gemini referrals
Google Analytics showing sessions referred from gemini.google.com and chatgpt.com, both trending upward month over month.

If the lines are going up, you're directionally right.

There are real attribution challenges with this approach. The much better method is to combine server log impressions with post-purchase surveys ("How did you hear about us?" with "AI search / ChatGPT" as an option). But referral tracking is free, it's immediate, and it tells you if what you're doing is working.

What's Next

There is a lot more you can do at every step. On-site content optimization, technical SEO for AI crawlers, off-site strategies, more advanced analytics. But what you've read here is a simple, do-it-yourself approach that is extremely cheap.

In fact, it's free right now.

and tell us what you find.