Follow the Money

Follow the Money

Momentic Research · May 2026

The AEO and GEO gold rush is full of people selling maps to new territory. A field guide to the incentives behind the tools and the claims, so you can tell signal from sales and keep the parts that work.

Most of the advice is sold by someone who profits from it

Why this exists

A new search surface arrived, and an industry rushed in to sell tools, dashboards, and retainers for it. Most of it is well meant and some of it is genuinely good. Almost all of it carries an incentive. The useful skill right now is reading that incentive without sliding into cynicism: you do not need to distrust everyone, you need to know who profits from each claim and ask one good question before you buy.

This is not a takedown. The mechanisms behind AI search are real, and the best tools earn their keep. The point is a working filter, so your budget goes to what moves the business and skips what only moves a dashboard.

Citation counts are being sold as a scoreboard

The most common pitch right now is a number that goes up: your citation share, your mention count, your AI visibility score. It feels like progress. Walked into a CFO meeting, it lands with a shrug. A rising mention count is the new version of impressions, motion that has not yet been shown to be money.,"Citation share is going up. Your CFO is going to yawn, or laugh in your face.",The retire list this year: domain authority, impressions, rankings, and now raw citation counts, unless each one is tied to an outcome. A metric with no dollar behind it is a story you tell yourself.

Everyone in the market wants something

The landscape · tap a category

Four kinds of players sell into AI search, and each has a built-in incentive. Knowing the incentive does not make them villains. It tells you which question to ask before you sign.

None of these are reasons to walk away. They are reasons to ask the question next to each one, and to keep the tools that answer it well.

• Visibility trackers — Dashboards that count your mentions and citations. Incentive: counting is the product, so the number has to look like the scoreboard. Ask: does this move pipeline, or just the dashboard?
• AI content tools — Fast pages and posts at scale. Incentive: more output looks like more value, whether or not it actually helps. Ask: does scaling this make me more distinct, or blend me into the crowd?
• GEO & AEO services — A new acronym, and a retainer to match. Incentive: novelty justifies the fee; a fresh label sells better than old work. Ask: is this a genuinely new discipline, or last year's work relabeled?
• The engines themselves — The surface, and the metrics it defines. Incentive: if they name the metric, success gets measured on their terms. Ask: who decided this number mattered, and what do they gain if I chase it?

You cannot trace a citation to a dollar

The honest part

Here is what the dashboards gloss over: nobody can yet draw a clean line from one AI mention to one sale. At SEO Week, the SEO equation was set next to the AI one, and every box on the AI side read "no idea" — prompt volume, click-through, conversion, all unknown. That does not make AI search worthless. It makes the confident ROI charts a story.

Treat AI presence as a leading indicator you watch over time, not a sale you can book. Anyone promising attribution down to the dollar is selling the one thing the data cannot give yet.

Source: Wil Reynolds, SEO math versus GEO math, SEO Week 2026.

Translate the metric, or stop reporting it

Make it count · pick a metric

The fix for a vanity number is not to ban it. It is to walk it down to a dollar. The move fits in one sentence: rate times volume times value times win rate equals pipeline.

If you cannot ladder a number down to revenue, it is a diagnostic at best and theater at worst. Report the ladder, not the raw count.

• Citation share — Vanity: your share of mentions went up; a CFO has no idea what that buys. Translated: 10% mention rate × 1,200 monthly questions × $500K deal × 20% win rate = pipeline you can say in one sentence.
• Mentions — Vanity: you were named more this month; so were a dozen competitors. Translated: mentions → branded search → sessions → leads → closed revenue = a tracked path, not a vanity count.
• Rankings — Vanity: you rank for more terms; rankings stopped equalling clicks years ago. Translated: visibility → assisted conversions in GA4 → revenue influence = a contribution you can defend.

Source: Dale Bertrand, prioritizing investment by financial impact, SEO Week 2026.

Judge the tool that ships, not the one announced

Announced vs shipped

The loudest moments in this space are launches. A suite of tools can be promised from the main stage and, months later, still not have shipped. No bad faith required — ambition outran delivery, which is ordinary. The lesson is for the buyer: weigh what is in your hands today, not what was promised from a stage.

Discount roadmap claims to near zero. The evidence that counts is a tool you can run, a dataset you can open, and a result you can reproduce yourself.

Every claim has a primary source. Find it.

Trace it back · tap a claim

The fastest way to test a scary or exciting statistic is to trace it to where it actually came from. Vague attribution is the tell: "studies show," "a leading analysis," "experts agree." Follow each one to the primary and the picture usually shrinks.

When a claim cannot survive the trace, drop it. When it can, you have something real to build on, and a source you can cite by name.

• "Studies show AI traffic converts at double digits." Trace it: a single vendor blog post, small sample, with a product to sell. The number traveled a lot further than its evidence.
• "A leading analysis found a 400% lift." Trace it: another company's dataset, relabeled by whoever reshared it. The original primary usually said something narrower and quieter.
• "Experts agree schema is the key to citations." Trace it: no named expert, just a roundup of opinions. Find one person willing to put their name and their method on it.
• "Our index shows you in 60% of answers." Trace it: their index, their prompts, their definition of an answer. Ask what was measured before you trust the share.

A field kit for the gold rush

What to do Monday

Five habits that turn skepticism into a tool instead of a posture, so you can spend on what works and ignore the noise.

• Follow the incentive. For any tool or claim, name who profits and what they want you to believe. Then weigh it on its merits.
• Run the dollar test. Make every metric ladder down to revenue. If it cannot, it is a diagnostic, not a result.
• Trace to the primary. Chase every "studies show" to a named source and method. Most claims shrink on contact.
• Buy what ships. Pay for tools you can run today. Discount roadmaps and stage promises to near zero.
• Keep the good stuff. Plenty here works. The filter exists to fund the real mechanisms, not to dismiss them.

Keep the signal. Skip the sales.

Where this leaves you

The mechanisms behind AI search are real, and so is the opportunity. The noise is just louder than usual, because the territory is new and the maps are for sale. Follow the money, ask the one good question, and you will fund the work that moves the business while the hype cycles past.