AI-verifiable client facts
Structure client identity, locations, products, services, offers, packages, prices, policies, availability, and customer action paths.
Commerce Graph holds what a client sells, where it is available, what the rules are, and what customers can do next. AI gets the approved version instead of guessing.
Each part has one job: make client facts, buyer paths, or checks clear enough for AI and people to use.
Structure client identity, locations, products, services, offers, packages, prices, policies, availability, and customer action paths.
Review, approve, publish, compare, and roll back graph versions so AI systems only receive the client facts you trust.
Attach claims to website pages, menus, listings, partner feeds, APIs, PDFs, or approved third-party sources so AI can verify what is true.
Keep internal notes and partner-only details separate from approved public output while still giving AI a clean way to read the client.
Start with the client facts as they are. Turn them into approved data, clear customer paths, checks, and fixes.
Import the website, menus, PDFs, public listings, partner feeds, and approved internal notes.
Map facts into business, location, offer, policy, source, and action fields with owners and confidence.
Review changes, resolve conflicts, publish a version, and expose only approved fields to endpoints.
Use endpoint checks and Answer QA to find stale facts, missing fields, and source conflicts.
The inputs and outputs stay visible, so the agency, client, and AI systems can use the same approved facts.
| Area | Input | Output |
|---|---|---|
| Client identity | Legal name, display name, categories, service regions, contact rules | Agent-readable client profile |
| Offers and packages | Menus, service descriptions, durations, prices, deposits, availability notes | Structured offer catalog |
| Policies | Cancellation, refund, deposit, arrival, group-size, eligibility, and safety rules | Policy summaries attached to offers and actions |
| Commercial actions | Booking links, forms, phone numbers, WhatsApp paths, partner handoffs | Approved action paths for agents |
AI systems can rely on approved graph data instead of guessing from stale pages, listings, documents, or fragmented snippets.
Agencies and commerce partners can onboard clients from one verified client model instead of repeated spreadsheets and manual checks.
Operators can see which facts are missing, disputed, or waiting for approval before those gaps cost recommendations.
The practical questions agencies and partners ask before they put Akii to work for a client.
No. Agencies or approved client owners approve graph versions and field visibility before data is published through Agent Gateway outputs.
Yes. The model is designed for services with locations, packages, menus, policies, deposits, capacity, and human handoffs.
Published versions are checked through endpoint monitoring, source drift checks, and Answer QA defects with retests.
Structure what the business sells, publish approved outputs, and test whether AI can find the right facts and customer path.