I asked ChatGPT to recommend you [02/25/26]

Your competitors aren't beating you with better features. They're beating you with better frames.

Here's what AI taught me: when a buyer asks an LLM to recommend solutions in your space, the model doesn't compare feature lists. It defaults to whoever owns the category definition. The company that named the problem gets recommended. The one still explaining their differentiators gets buried in "other options to consider."

Features live inside frames. Frames determine which features matter. When you position as "better," you're accepting someone else's frame. When you position as "different," you're still playing their game with modified rules.

AI amplifies this gap because models trained on millions of documents will always default to the strongest pattern, the clearest claim, the most defensible frame.

If a buyer can't explain why you're different in one sentence, neither can Claude.

The Positioning Conviction Test

Start with your positioning hypothesis. Write it as a single sentence:

Use this AI prompt: "Here's my positioning: [your sentence]. Now analyze 10 competitor websites in [your space]. For each, extract their positioning if stated. Then tell me: What's different about mine? What's defensible? What sounds like everyone else?"

Review the AI's analysis. Refine your positioning based on two questions:

If the difference isn't clear or isn't defensible, revise and test again.

Why this works: Most positioning is too vague or too easy to copy. This workflow forces you to test differentiation against real market language.

When to apply it: When launching a new offer. When repositioning your business. When prospects call, they reveal confusion about what you do. Anytime positioning "feels off."

Claude Projects (claude.ai)

This is where I test positioning claims before they go public. Create a project. Feed it your website copy, sales decks, and recent case studies. Then prompt: "Based on these materials, explain what this company does and why someone would choose them over alternatives."

What it does well: exposes weak framing immediately. If Claude defaults to feature comparisons or generic category language, your positioning isn't distinct enough to survive algorithmic interpretation. It shows you how pattern-matching systems perceive your claim.

Where it breaks: it won't create your frame. It will only reflect what you've already built. If your materials are vague, Claude's summary will be vague. This is diagnostic, not generative.

When to use it: before launching new messaging, rebranding, or updating your homepage. After you complete the Positioning Conviction Test, treat Claude as the pattern-recognition stress test.

When to skip it: if you're still in the idea stage with no existing materials. You need something to test first.

A founder told me last month: "We're the Shopify for X."

I asked what problem Shopify actually solved that made them different.

Silence.

Then: "I guess we're just saying we're easy to use?"

That's not explaining what you do. That's just borrowing someone else's success and hoping people get it.

Here's what changed.

Two weeks later, they rewrote their homepage. Instead of the comparison, they said this: "Most platforms assume you have a developer. We assume you don't."

Same product. Completely different explanation. Demo requests doubled.

So what happened?

The first version made you work to understand them. You had to know what Shopify does, then guess how they're similar, then figure out if that matters to you.

The second version just told you the problem they solve. If you don't have a developer and need a platform, you know immediately this is for you.

That's the difference between saying you're like something else and explaining the specific problem you fix.

If you read this edition and thought "I need to actually test our positioning," download the framework. Don't just think about it.

The Positioning Conviction Test is a 3-question diagnostic that forces clarity. It's the same system I walked through in this newsletter, packaged as a visual workflow you can run yourself or with your team.

You'll get the flowchart that maps all three questions, the exact prompts to use with Claude for competitor analysis, and the decision tree that tells you whether your positioning is locked or needs work.

This is the filter that separates positioning that survives AI recommendation engines from positioning that gets buried in "other options to consider."

P.S. If you run the test and realize your positioning doesn't pass, take the free 3-minute quiz at www.miahorm.com/start to identify your #1 bottleneck. If you're struggling with positioning, clarity is likely what's holding you back.