{ "format": "hyperize/v1", "@id": "https://www.hyperize.ai/en/insights/articles/can-ai-agents-buy-from-your-brand-seven-prompts.json", "type": "Article", "pageType": "Article", "title": "Can AI agents buy from your brand? Ask them.", "alternativeHeadline": "7 prompts to test if AI agents can find, use, and buy from your brand", "url": "https://www.hyperize.ai/en/insights/articles/can-ai-agents-buy-from-your-brand-seven-prompts", "alternateLanguage": { "de": "https://www.hyperize.ai/de/insights/articles/can-ai-agents-buy-from-your-brand-seven-prompts" }, "inLanguage": "en", "author": { "name": "Marc Seefelder", "url": "https://www.hyperize.ai/en/about#marc" }, "publisher": { "@id": "https://www.hyperize.ai/#organization", "name": "Hyperize", "url": "https://www.hyperize.ai", "parentOrganization": "MING Labs" }, "datePublished": "2026-07-03", "dateModified": "2026-07-03", "nextReview": "2027-01-03", "evidenceTier": "proprietary", "confidence": "B", "summary": "Seven copy-paste prompts that test whether AI agents can find a brand, read its pages, complete its transaction, and cite its evidence. A self-serve AI agent readiness audit: the diagnosis Hyperize runs against DAX 40 brands, adapted for any AI assistant. The prompts sort into four gates: FIND (prompts 1, 2, 7), READ (prompt 3), TRANSACT (prompts 4, 5), BE CITED (prompt 6). One run is a smoke test, not a measurement; the instrumented version fixes tasks, repeats runs, scores per lane, and re-probes on a schedule.", "answer": "You can test whether AI agents can find, read, and transact with your brand using seven prompts in any AI assistant: an invisibility check, a memory check, a blank-shell check, a task check, an interception check, an evidence check, and a vocabulary check.", "hook": "Seven prompts and one hour tell you what a year of SEO reporting cannot: whether AI agents can find your brand, read your pages, and finish a purchase. We run these tests against the DAX 40, wave after wave [S1]. The pattern is stable: strong brands, blank pages, intercepted sales [S2][S3]. Below are the exact prompts, what a failing answer looks like, and why each failure costs revenue. No tool. No signup. Run them today.", "setup": { "name": "Context block", "instruction": "Paste once, before any test. Fill in every bracket.", "prompt": "You are testing how AI agents handle the brand [brand]. Context: website [URL]. Category: [category]. Market: [market]. The transaction a customer completes with us: [buy / get a quote / book / apply]. Our three main competitors: [A], [B], [C]. Hold this context for every test that follows. After each test, report three things: what you found, what you could not find or do, and where you went instead." }, "prompts": [ { "number": 1, "name": "The invisibility check", "gate": "FIND", "conditions": "Fresh chat, logged out or temporary session, web search on.", "prompt": "Search the web and answer: which [category] providers in [market] would you recommend to a customer right now, and why? List every source you used.", "readTheAnswer": "Named and cited: your pages are material for the answer. Named but not cited: borrowed visibility, other people's pages control what the engine says. Absent: for this question, you do not exist.", "failSoundsLike": "Here are five providers, and you are not one of them." }, { "number": 2, "name": "The memory check", "gate": "FIND", "conditions": "Fresh chat, web search OFF. Then repeat in a second fresh chat with search on and compare.", "prompt": "What do you know about [brand]? Answer from memory only, do not search the web. Then state how current you believe your picture is, and what you are unsure about.", "readTheAnswer": "The gap between the memory answer and the search answer is your dependence on retrieval. Stale memory plus weak retrieval means agents work from the fossil layer.", "failSoundsLike": "A confident description of your 2023 company." }, { "number": 3, "name": "The blank-shell check", "gate": "READ", "conditions": "Point it at your most important product page.", "prompt": "Fetch [product page URL] and list everything a customer could learn from it: products, prices, specs, availability, and the next step. Then list what you could not read or open.", "readTheAnswer": "GPTBot, ClaudeBot and PerplexityBot do not render JavaScript (Gemini is the measured exception). If the page builds itself in the browser, most agents see scaffolding, not content.", "failSoundsLike": "The page appears to contain mostly scripts." }, { "number": 4, "name": "The task check", "gate": "TRANSACT", "conditions": "Use an assistant with browsing. Give it the real transaction.", "prompt": "Act for a customer who wants to [buy / get a quote for / book] [product]. Start at [URL]. Get as close as you can to the point where only payment or a signature remains. Log every step: what you clicked, what you could not read, where you stopped, and what you would tell the customer to do instead.", "readTheAnswer": "Read the log like an accident report: where exactly did the agent stop? The final line, what it would tell the customer instead, is the sentence your customer actually receives.", "failSoundsLike": "I could not proceed past the configurator." }, { "number": 5, "name": "The interception check", "gate": "TRANSACT", "conditions": "Fresh chat. Do not mention your brand.", "prompt": "I want to [buy / insure / book] [product category]. Where exactly should I do that? Name the one place you would complete it, then the second-best option.", "readTheAnswer": "Your domain: you own the close. A portal, marketplace, or dealer: the agent has decided your surface is not the place to finish the job. This is Third-Party Interception.", "failSoundsLike": "I would complete this on [comparison portal]." }, { "number": 6, "name": "The evidence check", "gate": "BE CITED", "conditions": "Sources mandatory. Read the citations, not the prose.", "prompt": "Why should a customer choose [brand] over [competitor]? Cite a source for every claim you make.", "readTheAnswer": "Your pages cited: you control the argument. Trade press and forums: others control it. No sources: the engine is improvising and will improvise differently tomorrow.", "failSoundsLike": "Every claim cited to trade press, none to you." }, { "number": 7, "name": "The vocabulary check", "gate": "FIND", "conditions": "Fresh chat. Do NOT mention your brand.", "prompt": "Explain [your coined concept or product term]. Who is the authority on it, and which sources would you cite?", "readTheAnswer": "You cited: the term is yours, build on it. A competitor cited: the term is not yours, rename or prefix. Nobody cited: open ground.", "failSoundsLike": "Your concept, explained with a competitor as the source." } ], "gates": [ { "name": "FIND", "prompts": [ 1, 2, 7 ], "axis": "AI Visibility" }, { "name": "READ", "prompts": [ 3 ], "axis": "AI Usability" }, { "name": "TRANSACT", "prompts": [ 4, 5 ], "axis": "AI Usability" }, { "name": "BE CITED", "prompts": [ 6 ], "axis": "Evidence layer" } ], "weightingNote": "FIND is the AI Visibility axis; READ and TRANSACT are the AI Usability axis; BE CITED is the evidence layer that feeds both. In the published Agent Success Score, AI Usability carries 70 percent of the weight.", "limits": "One prompt is one engine, on one day, with one phrasing. Engines drift week to week; answers vary run to run. A single pass is a smoke test, not a measurement. The instrumented version fixes the tasks, repeats the runs, scores per lane, and re-probes on a schedule, because citations decay.", "sections": [ { "id": "section-01", "title": "Your SEO audit cannot see this." }, { "id": "section-02", "title": "Load the context first." }, { "id": "section-03", "title": "Prompt 1 · The invisibility check." }, { "id": "section-04", "title": "Prompt 2 · The memory check." }, { "id": "section-05", "title": "Prompt 3 · The blank-shell check." }, { "id": "section-06", "title": "Prompt 4 · The task check." }, { "id": "section-07", "title": "Prompt 5 · The interception check." }, { "id": "section-08", "title": "Prompt 6 · The evidence check." }, { "id": "section-09", "title": "Prompt 7 · The vocabulary check." }, { "id": "section-10", "title": "Seven answers, one map." }, { "id": "section-11", "title": "What the prompts cannot tell you." } ], "twoSentenceStandout": "A page an agent cannot read does not exist. A transaction an agent cannot finish belongs to the middleman.", "sources": [ { "id": "S1", "publisher": "Hyperize", "title": "DAX 40 Agent Success Index, measured by real agents, wave by wave", "date": "live", "url": "https://www.hyperize.ai/en/dax40-index", "supports": "The measurement program these prompts are distilled from, and the pattern the four-gates map describes.", "type": "internal" }, { "id": "S2", "publisher": "Hyperize", "title": "Can AI agents use German car configurators? (Wave: Automotive)", "date": "2026-05", "url": "https://www.hyperize.ai/en/insights/articles/four-german-cars-tested-by-ai-agents", "supports": "The blank-shell finding behind prompt 3: a coding agent found 227 interactive elements on the BMW configurator and could read none of them.", "type": "internal" }, { "id": "S3", "publisher": "Hyperize", "title": "Three German insurers, tested by AI agents (Wave: Insurance)", "date": "2026-06", "url": "https://www.hyperize.ai/en/insights/articles/three-german-insurers-tested-by-ai-agents", "supports": "The interception pattern behind prompt 5: brands get named while intermediaries own the close.", "type": "internal" }, { "id": "S4", "publisher": "Hyperize", "title": "The Proof Flywheel, public citation ledger", "date": "2026-06", "url": "https://www.hyperize.ai/en/insights/articles/the-proof-flywheel", "supports": "The citation evidence behind prompts 1, 6 and 7: a vocabulary-led win and a plain-language loss on the same page; the whitespace loss diagnosed as entity association, not crawling.", "type": "internal" }, { "id": "S5", "publisher": "Hyperize", "title": "AI agents built our websites (making-of)", "date": "2026-06", "url": "https://www.hyperize.ai/en/insights/articles/ai-agents-built-our-websites", "supports": "The verified crawler finding behind prompt 3: none of the major AI crawlers currently render JavaScript.", "type": "internal" }, { "id": "S6", "publisher": "Hyperize", "title": "Agent Success Score, methodology", "date": "live", "url": "https://www.hyperize.ai/en/methodology/agent-success-score", "supports": "The published axes and weighting behind the map: AI Usability carries 70 percent of the score.", "type": "internal" }, { "id": "S7", "publisher": "Hyperize", "title": "Is your brand ready for AI agents? (the Fossil Layer Test)", "date": "2026-03", "url": "https://www.hyperize.ai/en/insights/articles/brand-ready-for-ai-agents", "supports": "The frozen-memory model behind prompt 2: training data ages like sediment; retrieval is the living layer.", "type": "internal" } ], "relatedSurfaces": [ { "relationship": "mentions", "target": "Agent Surface", "atId": "https://www.hyperize.ai/en/methodology/agent-surface#defined-term-agent-surface", "url": "https://www.hyperize.ai/en/methodology/agent-surface", "note": "The layer the prompts probe." }, { "relationship": "mentions", "target": "Third-Party Interception", "atId": "https://www.hyperize.ai/en/methodology/third-party-interception#defined-term-third-party-interception", "url": "https://www.hyperize.ai/en/methodology/third-party-interception", "note": "The pattern prompt 5 detects." }, { "relationship": "mentions", "target": "DAX 40 Agent Success Index", "atId": "https://www.hyperize.ai/en/dax40-index#dataset", "url": "https://www.hyperize.ai/en/dax40-index", "note": "The instrumented version of this diagnosis, running in public." }, { "relationship": "isPartOf", "target": "Insights", "atId": "https://www.hyperize.ai/en/insights#webpage", "url": "https://www.hyperize.ai/en/insights", "note": "Editorial layer of the Hyperize Agent Surface." } ], "scope": { "publishes": [ "Seven diagnosis prompts, verbatim, with reading instructions per prompt", "The four-gates model (FIND, READ, TRANSACT, BE CITED) and its mapping to the published score axes", "Already-published findings as anchors (blank-shell, interception, vocabulary effect, crawler-JS)", "Honest limits: one run is a smoke test, not a measurement" ], "doesNotPublish": [ "Build recipes (how to make a page pass the tests)", "Scoring derivation and task protocols", "Query-classification and trigger logic", "Measurement harness internals" ] }, "engagements": [ { "name": "Free Snapshot", "href": "/en/#snapshot", "role": "The instrumented version of this diagnosis: one lane, one task, measured and documented. Free." }, { "name": "DAX 40 Agent Success Index", "href": "/en/dax40-index", "role": "The diagnosis running in public, brand by brand, wave by wave." }, { "name": "Methodology", "href": "/en/methodology", "role": "Agent Success Score: two axes, published weights, per-lane measurement." } ] }