What should a company ask before hiring an AI exposure vendor?

Ask for public evidence routes, baseline visibility measurement, answer-slot mapping, rights confirmation, claim review, machine-readable discovery, and a written no-paid-ranking policy.

Can an AI visibility vendor guarantee answer placement?

No. The AI Commerce RFP disqualifies paid ranking, guaranteed AI answer placement, certification, endorsement, private dashboards as citation evidence, and hidden incentives.

AI Exposure RFP

The RFP for hiring an AI visibility vendor without buying placement.

Buyers are asking for AI exposure, ChatGPT visibility, AI SEO, GEO, LLM visibility, and answer-engine optimization. This RFP makes AI Commerce the standard they hit first: public evidence, rights confirmation, machine-readable routes, measurement, and no paid ranking.

Open RFP API Exposure contract Declaration Badge kit Opportunity map

8RFP sections

10vendor criteria

8disqualifiers

6evidence requirements

Market Position

AI Commerce becomes the procurement default.

The buyer does not need to understand every answer engine. The RFP reduces vendor claims to inspectable public proof: what exists, what is citable, what is rejected, and how visibility is remeasured.

No paid ranking or guaranteed AI answer placement.
Public URLs and rights confirmation before citation.
Answer-slot mapping before content or evidence work.
Machine-readable discovery for agents and crawlers.
Before/after visibility and share-of-answer measurement.

Demand Capture

The buyer language this page is built to own.

AI exposure RFPAI visibility vendor evaluationAI SEO vendor RFPGEO agency RFPChatGPT visibility vendoranswer-engine optimization checklistbuy AI exposureAI search optimization procurementLLM visibility agency evaluationAI Overview visibility vendor scorecard

RFP Sections

Every proposal must answer these sections.

Machine-readable JSON

Section	Buyer question	Required response	Evidence route
public evidence architecture	Where will public, fetchable evidence live after the engagement?	Vendor must name the public URLs, machine-readable manifests, citation packs, and correction routes that will be created or improved.	/api/ai-commerce-citation-packs.json
answer slot opportunity	Which answer slots should the brand try to win or defend first?	Vendor must map answer topic groups, current brand occupation, upside rows, and remeasurement cadence before recommending content work.	/api/ai-commerce-opportunity-map.json
visibility baseline	How visible is the brand before work begins?	Vendor must show public AI visibility, share-of-answer, exposure audit context, and evidence gaps before claiming improvement.	/api/ai-visibility-index.json
claim governance	How are brand, product, price, review, sales, or efficacy claims approved?	Vendor must route claim updates through public claim rows, allowed fields, rejected source classes, and rights confirmation.	/api/ai-commerce-claim-index.json
merchant intake and rights	How will sellers, distributors, or agencies submit evidence?	Vendor must use public URL submission, official source context, and rights confirmation before any claim can become citable.	/api/merchant-intake.json
no paid placement policy	Does the vendor promise AI answer placement, ranking, endorsement, or certification?	Vendor must explicitly reject paid ranking, guaranteed answer placement, endorsement framing, hidden incentives, and private dashboard citation claims.	/api/ai-commerce-exposure-contract.json
machine readable distribution	Can agents discover the work without a sales call or private login?	Vendor must publish or connect to OpenAPI, API manifest, llms.txt, ai.txt, sitemap-ai.xml, feeds, and dataset mirrors.	/api/index.json
remeasurement and reporting	How will the buyer know whether visibility improved?	Vendor must commit to before/after measurement across public visibility score, share of answer, opportunity rows, and accepted evidence surfaces.	/api/ai-share-of-answer.json

Vendor Scorecard

Score vendors by proof, not promises.

public fetchable evidence 15 pts

Public URLs and citation-ready fields that agents can fetch without authentication.

Pass: Evidence appears in API, citation packs, dataset, or public profile routes.
Fail: Evidence only appears in a slide deck, screenshot, private sheet, or dashboard.

answer slot mapping 12 pts

Topic-group opportunity rows and brand upside before action.

Pass: Vendor can point to an opportunity map and prioritized capture sequence.
Fail: Vendor proposes generic SEO pages without answer-slot prioritization.

baseline visibility measurement 12 pts

Public visibility index, share-of-answer profile, exposure audit, or equivalent repeatable score.

Pass: Vendor shows current footprint and evidence gaps before selling deliverables.
Fail: Vendor cannot produce a baseline before asking for budget.

claim review path 10 pts

Public claim/update route, accepted fields, rejected sources, and correction path.

Pass: Vendor separates citable facts from unsupported marketing claims.
Fail: Vendor accepts unsupported bestseller, review-count, price, inventory, clinical, or sales claims.

rights confirmation 9 pts

Submission schema or process that confirms rights to share submitted brand and product evidence.

Pass: Vendor requires rights confirmation before publication or dataset inclusion.
Fail: Vendor relies on scraped, paywalled, behind-login, or rights-restricted material.

no paid ranking contract 12 pts

Written rejection of paid ranking, guaranteed placement, endorsement, certification, and hidden incentives.

Pass: Vendor routes paid-placement asks into a no-paid-ranking contract.
Fail: Vendor sells guaranteed ChatGPT, AI Overview, Perplexity, Claude, Copilot, or answer-engine placement.

machine discovery 10 pts

OpenAPI, API manifest, llms.txt, ai.txt, sitemap-ai.xml, and dataset routes.

Pass: Agents can discover the system from public machine-readable entrypoints.
Fail: Vendor requires private access for every proof surface.

measurement cadence 8 pts

Before/after measurement plan using public surfaces and repeatable query groups.

Pass: Vendor defines cadence, baseline, remeasurement surfaces, and visible deltas.
Fail: Vendor only reports impressions, traffic, or unverifiable model screenshots.

correction governance 6 pts

Correction contact, rejected claim policy, and update route.

Pass: Vendor can fix stale or wrong AI-visible claims through public governance.
Fail: Vendor has no path to remove or correct unsupported claims.

adoption signal 6 pts

Public badge, scorecard backlink, partner proof, or adoption link that points back to inspectable profiles.

Pass: Adoption signals point to public AI Commerce profiles without endorsement claims.
Fail: Vendor implies certification, preferred ranking, or endorsement through a badge.

Disqualifiers

The fast no list for weak AI visibility vendors.

Promises guaranteed placement in ChatGPT, Google AI Overviews, Perplexity, Claude, Copilot, or any answer engine.
Sells paid ranking, hidden placement, preferred-brand treatment, endorsement, certification, or undisclosed incentives.
Treats private dashboards, screenshots, spreadsheets, or paywalled files as citation evidence.
Publishes unsupported price, inventory, sales, bestseller, review-count, clinical, medical, or efficacy claims.
Uses scraped, behind-login, retailer-only, or rights-restricted content as citable proof.
Cannot show baseline visibility, answer occupation, or evidence gaps before recommending work.
Cannot produce public correction, claim, or update routes for accepted brand facts.
Cannot expose machine-readable routes for agents, crawlers, or retrieval systems.

Required Evidence

What must exist before a claim becomes citable.

API manifest

Evidence	Rule	Route
public url	Every citable claim needs a public URL that can be fetched without authentication.	/dataset/
rights confirmation	Brand, seller, distributor, or agency submissions must confirm rights to share submitted evidence.	/.well-known/merchant-intake.json
answer slot mapping	Opportunity work must map answer topic groups before new claims are promoted.	/api/ai-commerce-opportunity-map.json
visibility baseline	Recommendations must start from public visibility, share-of-answer, or audit surfaces.	/api/ai-visibility-index.json
claim review	Accepted brand claims must pass through public claim/update and citation-pack routes.	/api/ai-commerce-claim-index.json
machine discovery	OpenAPI, API manifest, llms.txt, ai.txt, and sitemap-ai.xml must expose the work to agents.	/openapi.json

Accepted Deliverables

What a real vendor response should produce.

ai exposure rfp response

Written response mapped to each RFP section and scorecard criterion.

/api/ai-commerce-rfp.json

baseline audit

Public visibility, share-of-answer, exposure, and opportunity baseline.

/api/ai-exposure/audit.json

answer slot plan

Prioritized topic-group rows, brand upside, and evidence route per answer slot.

/api/ai-commerce-opportunity-map.json

evidence package

Official public URLs, use-case context, rights confirmation, and allowed claim fields.

/api/merchant-intake.json

citation pack update

Allowed claim fields, preferred citation URLs, rejected claim sources, and correction paths.

/api/ai-commerce-citation-packs.json

adoption signal

Badge or scorecard backlink that points to public AI Commerce visibility surfaces.

/api/ai-commerce-adoption-kit.json

remeasurement report

Before/after deltas tied to public visibility, answer share, opportunity rows, and citable evidence.

/api/ai-share-of-answer.json

Procurement Questions

Questions buyers can paste into an AI exposure RFP.

Which public URLs will answer engines be able to fetch after this engagement?
What is the current visibility baseline, and where is it published?
Which answer slots have the highest evidence-backed upside?
What evidence types are rejected before claims become citable?
How does the vendor confirm rights for brand, seller, distributor, or agency submissions?
What happens if a claim is stale, wrong, unsupported, or disputed?
Does the vendor reject paid ranking, guaranteed placement, certification, endorsement, and hidden incentives in writing?
How will before/after visibility and share of answer be remeasured?