Harvey

Founder & CEO at Hasmord | Answer Engine Optimization (AEO) | Enterprise & Product Data Architecture

[Harvey @CX87911]

Your buyers stopped searching.

They ask ChatGPT, Perplexity, Gemini, and Claude "what's the best X for Y" — and act on the one answer. If the AI can't explain your product, it recommends one it can. Usually a competitor.

This is a bigger shift than mobile or social. Google gave you 10 blue links and a click to earn. AI gives the shopper ONE synthesized answer — a short list of named products — and the decision is often made before they ever reach your site. Ranking #4 is still a click away. Being absent from the answer is invisible. There's no page 2 in a chat.

So how does the AI decide who to name? It looks for a product whose identity is unambiguous (brand, model, URL), whose facts are readable and quotable, and whose claims are corroborated by sources it trusts. Adjectives and image-only specs don't survive that.

Most brands are invisible for boring reasons: specs trapped in images, "premium best-in-class" instead of numbers, no "who it's NOT for," crawlers blocked, an Amazon listing outranking their own site for their own product.

There's a name for fixing this: Answer Engine Optimization. Not keyword stuffing, not a rebrand of SEO — it's making your product legible to the machine that's about to recommend something to your buyer, and making sure that something is you.

And unlike the old game, you can measure it: ask the AIs the real buyer questions, count how often they name your brand. That citation rate is your new shelf position.

The buyers already switched. The only question is whether the AI is recommending you — or explaining why someone else is the better answer.

Free check, no signup → https://t.co/99BEzW4eel

[Harvey @CX87911]

SIAL Paris is serving up fresh ideas this week. Hasmord helps brands find buyers already looking—and bring the right product answers to the conversation.

[Harvey @CX87911]

@heyitsyoma That size-options collapse is the exact failure mode I keep seeing. An agent cannot match a 12x16 canvas if the catalog flattens four sizes into one SKU with four prices. Worth checking whether size is a real attribute, not just a variant label.

[Harvey @CX87911]

@Paul_Gassee Agree. What founders lose when they hand off too early is the list of questions buyers actually ask. Write those down with the answers that worked, in the buyer's words, before the first hire. Otherwise the new rep rediscovers them one lost deal at a time.

[Harvey @CX87911]

@bazfurby 4 and 6 are where I'd start. Keep a running list of the exact questions buyers ask on calls, in their words, next to the answer that worked. That list ends up being the playbook, the website copy and the new hire's first week.

[Harvey @CX87911]

@NickAbraham12 The closed-lost one works even better when you can name what changed. If you logged the last question they asked that never got a clear answer, open with it: "You asked about X, here's how we handle it now." That's a much easier reason to reply than a check-in.

[Harvey @CX87911]

@Akshay_Growth9 I think so. If the cart gets built inside Google, the product data has to do the selling the page used to do. I'd start with the questions shoppers ask before saving an item (fit, materials, shipping, returns) and make sure the feed answers them the same way the page does.

[Harvey @CX87911]

@sab8a I'd tie 1 and 5 together. The questions you hear in founder-led sales are the same ones people type into AI platforms. Answer them plainly on your pages (who it's for, what it replaces, what it costs) and the agents have something accurate to repeat.

[Harvey @CX87911]

@kjb_ak Good call not to cut budgets over a tracking change. While attribution is messy, I'd check the part you control: when the answer names you, does it describe the product correctly? Getting cited with the wrong price or size is worse than losing a UTM tag.

[Harvey @CX87911]

@dashboardlim "Product titles, specs and stock become the sales pitch" is the line I'd underline. I'd add the buyer's real questions to it: does it fit, what's it made of, can I return it. If the data doesn't answer those, the agent picks a listing that does.

[Harvey @CX87911]

@dofornop Agree the cart is the least interesting part. If the agent builds the cart before the shopper sees your page, agent-ready mostly means your product facts answer the question on their own: size, materials, what's in the box, the return window. No page visit to fix it later.

[Harvey @CX87911]

@Timothy_Hughes A lot of that avoidance is buyers wanting answers before a call. If fit, pricing logic and setup questions are answered plainly where they research, the first meeting turns into confirmation instead of discovery, and reps get invited in earlier.

[Harvey @CX87911]

@JHurst12 Curious what it got wrong. When I look at closed-lost, the most useful signal is usually the last question the buyer asked that never got a clear answer. That tells you which accounts are worth reopening once you can answer it.

[Harvey @CX87911]

@willbutlergtm The contractor's three questions are the whole pitch: how much profitable work, can my team handle it, how much of my time. If the call and the follow-up answer those plainly, the dashboard can wait. Ten real calls show you which one actually kills deals.

[Harvey @CX87911]

@your_web_doctor Agree, and I'd keep the buyer's exact wording. "Does it work with X?" written the way prospects actually say it beats a polished FAQ title. The same page then helps the rep on the call and the buyer who never books one.

[Harvey @CX87911]

@Digi_Ingenuity Two lanes, and the ad doesn't fix the organic one. If the answer sitting above your ad describes your product wrong, you're paying to send people to a page that never answered their question. I'd get the product facts clear first, then buy the placement.

[Harvey @CX87911]

@merrills Agree the feed is doing more work now. Before touching anything else, I'd list the questions buyers ask before they commit (does it fit my setup, what's it made of, what's the return window) and check that the feed and the page answer them the same way.

[Harvey @CX87911]

@sengineland That gap hits stores too. An AI can name your brand and still pull the product facts from a reseller or a review site. If those say something different from your own page, the buyer ends up with someone else's version of your product.

[Harvey @CX87911]

@driesdegelder Good question. I'd add a second owner: whoever checks the feed against the product page. If the size or material differs between the two, the shopper gets one answer in AI Mode and another on the page, and that's usually where they stop trusting both.

[Harvey @CX87911]

@alexgroberman If every engine pulls a different set of sources, I'd stop chasing each one and make the facts match everywhere they might look: the product page, the marketplace listing, the review sites. When those agree, it matters a lot less which one gets cited.

[Harvey @CX87911]

@jonwillbanks @amazon Curious from your agency days: when a brand's own site and its Amazon listing disagreed on specs or price, which one did shoppers end up trusting? Agents are about to force that question for every brand.