AEO and sales
You can measure more of it than most dashboards show, but not all of it. What AI answers say about your products and the AI referral visits your analytics can see are measurable today. Leads, deals and revenue need your sales data and one simple habit: asking buyers where their research happened. This guide is for marketing, ecommerce and sales leaders who need numbers they can defend. If you only want a chart of brand mentions, it goes further than you need.
Last updated September 29, 2026 · ~7 min read
Buyers ask an assistant, read the answer, and then often buy somewhere else: on a marketplace, from a dealer, in a store or through a sales rep. Many never click a link at all. Pew Research Center found that when Google showed an AI summary, users clicked a traditional search result in 8% of those searches, against 15% when there was no summary. So the most visible number, AI referral traffic, is also the most incomplete. Measuring AEO means following the buyer through several layers, each with its own evidence.
| Layer | What it tells you | How to measure it today | Watch out for |
|---|---|---|---|
| 1. Answer | Whether AI answers name your product, describe it correctly and compare it fairly for the questions buyers ask | Run the same buyer questions through AI models on a schedule and keep the answer text, model and date | Model APIs can answer differently from consumer apps; one answer is not a pattern |
| 2. Visit | Whether answers send buyers to your pages | AI referral visits in your analytics, by landing page and product | No-click journeys, buying off-site and visits with no referrer are missing |
| 3. Lead or deal | Whether buyers act on what they learned | An AI question at intake, a field on the deal, and the buyer's own words when they mention it | Buyers rarely volunteer where they started; unknown is not the same as no |
| 4. Sale | Whether AI-influenced deals close, and for how much | Won and lost deals with their AI status, plus a short post-purchase question | Influenced is not caused; do not add contract, order and payment amounts together |
None of these layers stands in for the next. A rising mention count does not show that deals followed, and a flat referral chart does not show that AI stopped mattering.
The layer most teams skip is the one that matters most to sales: the deals where AI played a part. The rule that works is simple. If a buyer tells you an assistant helped them understand, compare or choose, count the deal as AI-influenced, even if there was never a tracked click and even if the order came through a dealer or a sales rep. A buyer's own statement, or a rep's record of what the buyer said, is evidence.
Report two numbers side by side:
Do not project a small sample onto every sale, and do not leave out confirmed deals because coverage is incomplete. The first overstates the effect; the second hides it.
An AI-influenced deal means AI took part in the buyer's research. It does not mean AI created the whole sale, and the deal value should never be presented as revenue generated by AEO alone. If you want to estimate the extra sales a change produced, you need a comparison: the same questions before and after a fix, or changed products against similar unchanged ones over the same period. Even then, promotions, price changes and seasonality can move the result. Say so in the report, and keep influence, visits and estimated lift as separate numbers.
What stays unknown: how each assistant chooses what to name, and every conversation a buyer had before they reached you. A good report makes those limits visible instead of filling them with guesses.
Being named in an AI answer is one step, not the result. What matters is whether the buyer’s question was answered well enough to move them toward a purchase, and what happens next. Today you can follow AI answer checks and AI referral visits in Google Analytics 4; inquiries, deals and contracts feed back into the same loop (coming in v3.0). The questions buyers raise in sales conversations point back to the product information: add evidence, sharpen the claim, then see how the market and sales respond.
Hasmord covers the first two layers today. It runs buyer questions through the Anthropic, Google, OpenAI and Perplexity APIs and keeps the answer text with its model and date, so the same questions can be compared after a change. It reads AI referral visits from Google Analytics 4 by product. And it finds the product-information gaps behind buyer questions, drafts fixes your team approves, and lets approved changes be rolled back. Recording AI influence on leads, deals and contracts inside Hasmord, and feeding sales results back into the next round, is coming in v3.0; until then, keep that status in your CRM using the plan above. If you are comparing tools, see how to choose an AEO platform for brands.
Hasmord is the AEO-native sales platform. It finds the buying questions your products can’t yet answer across sales and AI shopping channels, turns them into channel-specific improvements your team reviews, and checks what changes afterward. Lead-to-contract sales management and sales feedback are coming in v3.0.
Partly, and more than most teams do. You can measure what AI answers say about your products, the AI referral visits your analytics can see, and the deals where buyers tell you AI helped them decide. You cannot see every conversation a buyer had with an assistant, so report confirmed impact together with how many buyers you asked, and keep the rest as unknown.
Because many buyers never click. They read the answer, then search your brand, visit a marketplace or call a dealer. Some assistant traffic also arrives without a referrer and looks like direct traffic. Treat AI referral visits as a floor, not the full effect.
Ask. Add a question at intake or after purchase that lists AI assistants next to your other channels, and record the answer on the deal. A buyer saying an assistant helped them compare or choose is enough to count the deal as AI-influenced, even without a tracked click. Keep unanswered deals as unknown rather than guessing.
Not necessarily. An AI-influenced deal means AI took part in the buyer's research, not that it created the whole sale. To estimate the extra sales a change produced, you need a comparison, such as the same buyer questions before and after a fix, or changed products against similar unchanged ones, and even then the result is an estimate.
Related: how to estimate AEO ROI before you invest · The State of AI Product Discovery 2026 · how to get cited by ChatGPT · a case study from buyer question to fix.