MOFU: Are You in the Comparison?

The buyer knows the category now. They have a few names in mind, and they are asking AI to line them up. This is where the long list of TOFU becomes a shortlist of three to five brands, and where AI starts explaining why one is different from another.

Miss this stage and it doesn't matter how good the product is. The buyer never compares you, so they never choose you.

What AI is doing at this stage

AI is narrowing and contrasting. It takes the buyer's use case, audience, and region, filters the category down to the brands that plausibly fit, and then describes how they differ: strengths, weaknesses, who each one is best for.

Two things decide who survives the cut:

1. Differentiation AI can articulate. If AI can't say what makes you different, it has no reason to include you over a brand it can describe.
2. Presence in comparison sources. Comparison pages, review platforms, and "X vs Y" content are where AI finds the contrasts it reports. If those sources compare your competitors to each other and leave you out, AI tends to do the same.

The prompts buyers actually type

These are the five MOFU templates SurfaceMap runs. Three name your brand and two do not, because being described well when asked and being included when not asked are different tests.

ID	Shape	Question	What it tests
M1	Named comparison, competitor 1	How does (Brand Name) compare to (main competitor 1) for (use case)?	How AI frames you against your primary rival, and who it favours
M2	Named comparison, competitor 2	How does (Brand Name) compare to (main competitor 2) for (use case)?	The same test against your second rival. Results often differ.
M3	Unbranded forced choice	For (Target Audience) in (Region), which is the better choice for (use case): (main competitor 1) or (main competitor 2)?	Whether AI brings you up on its own when the buyer is comparing your competitors. This is the sharpest test of real competitive presence.
M4	Unbranded needs-filtered directory	Which (Category) options in (Region) best support (use case) for (Target Audience)?	Whether you survive the first filter from category to fit
M5	Named vs. general field	How does (Brand Name) perform for (use case) compared to other (Category) options in (Region)?	Whether AI can place you relative to the field, not just one rival

M3 deserves attention. If buyers are comparing your two biggest competitors and AI never mentions you as a third option, you are not really in the consideration set, whatever the branded questions say.

Fan-out queries for this stage

MOFU fan-out is where competitor names start appearing in the searches AI runs. High-likelihood patterns:

• (Brand Name) vs (main competitor 1)
• (main competitor 1) vs (main competitor 2)
• (main competitor 1) alternatives
• (Brand Name) alternatives
• (Brand Name) reviews
• (Brand Name) pros and cons
• (Category) comparison (Region)
• (Category) for (use case)
• best (Category) for (Target Audience) (current year)

Note the "alternatives" queries. When AI searches for alternatives to your competitor, the pages it finds (often written by other competitors) decide whether you appear in the comparison. SurfaceMap shows you which of these fan-out queries each model actually ran and which domains answered them.

Where AI looks for answers

MOFU is the transition stage. Third-party sources still carry weight, but the questions are getting specific enough that coverage starts to thin.

Typical sources at this stage:

• Review platforms with side-by-side comparison features
• "X vs Y" and "alternatives to X" articles
• Competitors' comparison pages (often the most detailed ones available, and never neutral)
• Analyst and trade reports
• Forum threads where buyers compare options
• Your own feature, service, and use-case pages

Here is the pattern. When the buyer asks "X vs Y" with no other constraints, there are usually plenty of third-party pages to cite. Add "for (use case)" and "for (Target Audience) in (Region)" and the number of independent sources that match all three drops sharply. At that point AI starts pulling directly from vendor pages to fill in the gaps: what each brand offers, who it's for, how it's positioned.

That means your own site is now competing with your competitors' sites to describe you. If a competitor has a clear comparison page that frames you unfavourably and you have nothing that states your own differentiation, AI has only one version to work from.

The trend continues into BOFU, where the constraints stack further and your own content becomes the main source.

What failure looks like

These AI Mention States show up most often as a MOFU problem. See the 12 AI Mention States for the full framework.

• Research Casualty: named at TOFU, but dropped once the buyer starts researching options.
• Comparison Drop-Off: absent from the side-by-side, especially in unbranded questions like M3 and M4.
• Interchangeable Vendor: included, but described so generically that AI gives the buyer no reason to prefer you.

How to measure it

• Unbranded inclusion rate (M3, M4). The most honest read on whether you are actually in the consideration set.
• Framing in named comparisons (M1, M2, M5). Who does AI favour, and on what grounds? Are its stated weaknesses accurate?
• Differentiator recall. Does AI mention your key features, or describe you in category-generic terms?
• Cited domains in comparisons. Whose comparison pages is AI reading? Are any of them yours?
• Model-by-model differences. A brand can make Perplexity's shortlist and miss ChatGPT's. The fix depends on which sources each one leans on.

How to fix it

On your own site

• Publish honest comparison pages for your main competitors. If you don't frame the comparison, someone else will.
• State your differentiators explicitly and specifically. "Great service" is not a differentiator AI can repeat. "Four-hour on-site response in (Region)" is.
• Build use-case pages that connect a specific buyer and problem to your offer, so you survive the M4-style filter.
• Mark up products, services, and use cases with structured data so the facts AI needs for a comparison are easy to extract.

Off your site

• Get profiles on the comparison and review platforms appearing in your cited domains, complete with the same differentiators.
• Earn inclusion in the "alternatives to (competitor)" content AI is reading.
• Correct inaccurate claims about you on third-party comparison pages where you can.

See where you stand at MOFU

• Run a free audit to see whether AI includes you when buyers compare your competitors, and how it frames you when they compare you directly.
• Take the 2-minute diagnostic to find your AI Mention State.

Previous stage: TOFU · Next stage: BOFU: Does AI recommend you, or the competitor?

FAQ

What is MOFU in AI visibility? MOFU (middle of funnel) is the Consideration stage of the AI buyer journey, where AI narrows a category down to a shortlist of three to five brands and compares them for the buyer's use case.

Why does AI compare my competitors but leave me out? Usually because the comparison sources AI reads (review platforms, "X vs Y" articles, alternatives pages) include your competitors and not you, and your own site doesn't state your differentiation clearly enough for AI to add you on its own.

Should I publish comparison pages about my competitors? Yes, if they are accurate and fair. At the consideration stage AI increasingly relies on vendor pages to fill gaps, and a clear, honest comparison gives it your framing to work from instead of a competitor's.