generated: '2026-09-19'
method: derived
source: >-
  json-schema/aircover-pipeline-*.json (Aircover's own published draft-2020-12 schemas from
  github.com/Aircover/aircover-pipeline) + the documented MCP tool arguments in mcp/aircover-mcp.yml, whose
  evidence is the provider's open-source Agent Skills. NOT derived from openapi/aircover-openapi.yml — that
  contract models only OAuth and the JSON-RPC transport and contains no domain entities.
confidence: medium
caveat: >-
  Aircover publishes no object reference and no schema for the MCP tool responses. The entities and fields
  below are the ones its own published artifacts name; field lists are partial by construction and types are
  only asserted where a JSON Schema states them. Nothing here is inferred from a response body, because the
  live surface is auth-gated.
entities:
- name: Meeting
  id_field: id
  id_shape: 32-character opaque string (the skills insist on the "full 32-char id")
  fields:
  - name: id
    type: string
    source: json-schema/aircover-pipeline-meetings-1.0.json
  - name: deal_id
    type: string
    note: "In the pipeline schema this field carries the COMPOSITE '<prospect-domain>/<deal-id>'; empty string when the meeting has no linked Salesforce deal. The MCP tools call that composite deal_key and reserve deal_id for the bare integer."
    source: json-schema/aircover-pipeline-meetings-1.0.json
  - name: team_ids
    type: string (comma-separated)
    source: json-schema/aircover-pipeline-meetings-1.0.json
  - name: date
    type: string (YYYY-MM-DD)
    source: json-schema/aircover-pipeline-meetings-1.0.json
  - name: notes_sent_to
    type: string (comma-separated emails of the deal team)
    source: json-schema/aircover-pipeline-meetings-1.0.json
  - name: prospect_org
    type: string
    source: mcp tool responses (aircover-skills)
  - name: start_time
    type: string (ISO timestamp)
    source: mcp tool responses (aircover-skills)
  - name: deal_team
    type: array
    source: get_meeting (aircover-skills)
  - name: prospect_attendees
    type: array
    source: get_meeting (aircover-skills)
  - name: notes
    type: object (AI summary)
    source: get_meeting (aircover-skills)
  read_by: [list_meetings, get_meeting]
- name: Transcript
  id_field: meeting_id
  fields: []
  note: Fetched whole by get_meeting_transcript(meeting_id); no published shape. Meetings without a transcript exist and cause agent_results to error.
  read_by: [get_meeting_transcript]
- name: Deal
  id_field: deal_key
  id_shape: '"<prospect_org>/<deal_id>" — the skills state the bare integer deal_id "resets" and must never be used alone'
  fields:
  - name: prospect_org
    type: string
  - name: deal_id
    type: integer
  - name: deal_key
    type: string
  - name: nextMeetingDate
    type: date
    source: get_deals (aircover-skills)
  - name: crm
    type: object
    note: Mirrors Salesforce and HubSpot fields — sfdc_opportunity_name, sfdc_opportunity_id, amount, close date, stage. An empty crm block is a placeholder, not an error.
    source: get_deal (aircover-skills)
  - name: completed_meetings
    type: array of Meeting
    source: get_deal (aircover-skills)
  read_by: [get_deals, get_deal]
- name: Agent
  id_field: id
  alias: template
  fields:
  - name: id
    type: string
    note: Passed as template_id to agent_results and get_qualification_results.
  - name: name
    type: string
  - name: category
    type: string
    note: 'e.g. "qualification"'
  types:
  - Extraction agents (MaxScore = 0) — pull structured data, no scoring
  - Scoring agents (MaxScore > 0) — evaluate against MEDDPICC, BANT, SPIN, Sandler and others
  - Hybrid agents — extraction plus a summarizing property
  - Dispatch agents — weekly report data and customer summaries
  source: https://www.aircover.ai/llms.txt
  read_by: [list_agents]
- name: AgentResult
  id_field: (meeting_id, template_id)
  fields:
  - name: meeting_id
    type: string
    source: json-schema/aircover-pipeline-agent-output-1.0.json
  - name: deal_id
    type: string
    source: json-schema/aircover-pipeline-agent-output-1.0.json
  - name: date
    type: string
    source: json-schema/aircover-pipeline-agent-output-1.0.json
  - name: template_id
    type: string
    source: json-schema/aircover-pipeline-agent-output-1.0.json
  - name: properties
    type: array
    note: Each property carries a title, a result and (for scoring agents) a score and max score. Join on title, NEVER on the positional key — the provider's skills say keys are positional and vary.
    source: json-schema/aircover-pipeline-agent-output-1.0.json
  read_by: [agent_results]
- name: QualificationResult
  id_field: (deal_key, template_id)
  fields: []
  note: Deal-level qualification scoring (dimension, score, max_score, result_summary per the path-to-quota skill's CSV columns).
  read_by: [get_qualification_results]
- name: Team
  id_field: id
  fields: []
  read_by: [get_teams]
  note: Referenced from Meeting.team_ids.
- name: Report
  id_field: id
  fields: []
  read_by: [list_reports, read_report]
  note: Weekly reports produced by dispatch agents.
- name: Document
  id_field: null
  fields: []
  note: Indexed content from Google Drive, OneDrive or Highspot; queried by natural language rather than by id.
  read_by: [ask_documents]
relationships:
- from: Deal
  to: Meeting
  type: has_many
  via: completed_meetings
- from: Meeting
  to: Deal
  type: belongs_to
  via: deal_id / deal_key (prospect_org + deal_id)
- from: Meeting
  to: Team
  type: has_many
  via: team_ids
- from: Meeting
  to: Transcript
  type: has_one
  via: meeting_id
- from: Meeting
  to: AgentResult
  type: has_many
  via: meeting_id
- from: Agent
  to: AgentResult
  type: has_many
  via: template_id
- from: Deal
  to: QualificationResult
  type: has_many
  via: deal_key