Back to About

About · Data model

What the platform knows about

22 entity types · 171 seeded records · click any entity to see the surfaces that render it.

Architecture diagram

Entity types

22

Distinct domain entities

Total records

171

Seeded across the prototype

Cross-references

87

Foreign-key edges between entities

AI surfaces grounded

28

Citations resolve into this graph

Three-tier hierarchy

Company · Client · Event — the spine of the data model.

Entities

22 types · 171 records
Client
Core

2 records

Fields

idnameshortNamelogoUrl
Event
Core

7 records

Fields

slugnameclientstatusstartDatebudgetregion
Vendor
Core

14 records

Fields

slugnamecategoryspendYTDriskFlagscontract
Contractor
Core

29 records

Fields

slugnameroledayRatecurrentlyAllocatedPctregions
Venue
Core

8 records

Fields

slugnamelocationcapacitiesrestrictions
Contract
Core

12 records

Fields

slugvendorSlugeventSlugstatusanomalies
EventTemplate
Core

6 records

Fields

slugnameeventTypephasesfields
BriefTemplate
Core

3 records

Fields

slugeventTypefieldsusageCount
JobCode
Resourcing

15 records

Fields

codeeventSlugpersonIdbudgetedHoursactualHoursToDateweeklyHours
ProjectBurn
Resourcing

5 records

Fields

eventSlugforecastHourshotFlagvelocityTrend
ContractorFeedback
Resourcing

10 records

Fields

contractorSlugeventSlugratingOverallbehaviorNotes
RetainerEngagement
Resourcing

1 record

Fields

clientIdtotalHoursmonthlyReviewsretainerLineItems
VendorProposal
Documents

7 records

Fields

vendorSlugeventSlugextractedItemsgapsclarifyingQuestions
CatalogItem
Documents

30 records

Fields

idnamevendorSlugunitPriceapprovedForClient
MoodBoardSearch
Documents

2 records

Fields

idquerytopMatchesoutreachDraft
PitchDeck
Documents

1 record

Fields

slugclientslides

Rendered on

RFPInbound
Documents

1 record

Fields

slugprospectbriefSummarycapabilityScores

Rendered on

VenueRegulations
Operations

3 records

Fields

venueSlugunionLaborriggingmandatoryVendors

Rendered on

GeoCostScenario
Operations

3 records

Fields

iditemCategorydestinationoptions
IntegrationConnector
Operations

9 records

Fields

slugnamestatuslastSync
VizIDPackage
AI

2 records

Fields

clientIdseriesNamemoodTilescolorPalette
MeetingTranscriptFixture
AI

1 record

Fields

slugattendeesdecisionsactionItemsquestions

Rendered on

Cross-references

How entities link to each other. The Playbook’s graph is dense — the AI uses these edges as citation paths.

  • EventVendor
    via Contract / VendorProposal / EngagedVendoran event lists its engaged vendors; each engagement traces back to a contract and a proposal
  • EventContractor
    via JobCodejob codes attribute hours to a person against a specific event
  • EventVenue
    via event.venue fieldone venue per event; venue regulations cascade to the event
  • EventEventTemplate
    via event.template fieldtemplates seed the BOM, phases, and budget skeleton
  • EventProjectBurn
    via eventSlugforecast vs actual hours rolled up per event
  • VendorContract
    via vendorSlug + eventSlugevery contract links a vendor to an event with anomaly findings
  • VendorCatalogItem
    via vendorSlugreusable items roll up to the vendor that supplies them
  • ContractorContractorFeedback
    via contractorSlug + eventSlugevery past event with a contractor has a behavior + delivery rating
  • VenueVenueRegulations
    via venueSlugunion labor, rigging caps, and mandatory vendors travel with the venue
  • ClientVizIDPackage
    via clientIdeach client has one brand kit — series names, mood tiles, palette
  • ClientRetainerEngagement
    via clientIdmonthly hours, reviews, and retainer line items per client
What the AI sees

When you ask the Playbook a question, it walks the graph: an event leads to its vendors and contractors, a vendor leads to its proposals and contracts, a contractor leads to feedback and job codes. Every answer cites a path through this graph — never an opinion, always an entity, a field, and a value.

The AI flags; the producer decides. Citations resolve into the same entity surfaces a human producer would open by hand — so every number is auditable, every claim is traceable, and no answer is a black box.

AI grounded in 171 seeded records across 22 entity types.