ChatGPT's Operator Layer Now Routes Purchases: Schema Decides Who Wins
OpenAI's operator API now supports end-to-end purchase routing through ChatGPT agents, making structured product and pricing data the deciding factor in which businesses get transacted with, not just cited. If your schema isn't machine-readable and verified, you're invisible to the layer that's now executing real buying decisions.
This Week's Signals
OpenAI Operator API Enables In-Agent Purchase Routing, Shifting Commerce to Schema Layer
Why it matters for your score
OpenAI's operator framework now allows third-party businesses to embed transactable surfaces inside ChatGPT agent flows, meaning an AI agent can discover, evaluate, and complete a purchase without the user ever visiting a website. Businesses without JSON-LD Product and Offer schema that ChatGPT can parse at crawl time are structurally excluded from these flows. This isn't a future risk; operator integrations are live and expanding across e-commerce and services verticals.
Google's AI Overviews Citation Study: Pages With HowTo and FAQPage Schema Cited 3.1x More Often
Why it matters for your score
A July 2026 analysis of 10,000 AI Overview citations found that pages deploying HowTo, FAQPage, or Speakable schema were cited 3.1 times more frequently than equivalent-authority pages without structured markup. The mechanism is direct: AI Overviews pull structured answer fragments rather than running full semantic inference when schema is present, making schema a citation shortcut. For local services and professional firms, FAQPage and Speakable markup are now measurably conversion-linked, not just an SEO hygiene item.
llms.txt Working Group Publishes Draft Verification Spec: Unsigned Files May Be Deprioritized
Why it matters for your score
The informal llms.txt working group released a draft verification specification in July 2026 proposing cryptographic signing of llms.txt files to let AI crawlers distinguish authoritative business-published summaries from scraped or hallucinated variants. Claude and Perplexity have both indicated interest in weighting signed files more heavily in grounding pipelines. Businesses that deploy llms.txt now, before signing is required, will be better positioned to upgrade in place rather than scramble when signing becomes a de facto standard for crawler trust.
What to do this week
- 1
Audit your JSON-LD Product and Offer schema this week using Faro's AI Schema tool (/tools/ai-schema), specifically check that 'price', 'priceCurrency', 'availability', and 'seller' fields are present and machine-readable. OpenAI's operator routing depends on these fields being parseable at crawl time, and missing even one required property can drop you from transactable results.
- 2
Add or update FAQPage and Speakable schema to your top service and landing pages, the July citation study shows this is the highest-leverage structured data move for AI Overview inclusion right now. Use Faro's AI Schema generator (/tools/ai-schema) to produce valid markup, then validate against Google's Rich Results Test before deploying.
- 3
Deploy or refresh your llms.txt file before the verification spec firms up; use Faro's llms.txt generator (/tools/llms-txt) to produce a clean, structured summary of your business, services, pricing, and coverage area. Getting a well-formed file indexed now means you'll only need to add a signature layer when the spec is finalized, rather than starting from scratch under time pressure.