curl --request POST \
--url https://go.aiinsurance.io/api/v1/companies/{companyId}/imports/jobs \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data @- <<EOF
{
"entityType": "Exposure",
"source": {
"importBatch": "october-load"
},
"transformations": {
"exposureName": "row.name",
"exposureType": "'company'"
},
"guards": [
{
"condition": "exposureName != null",
"errorMessage": "A name is required"
}
],
"attemptBatch": "mapping-v2"
}
EOF{
"id": "550e8400-e29b-41d4-a716-446655440003"
}Queue Import Job
Maps stored import rows into records in bulk, in the background, for a
migration or spreadsheet load. Store the rows first with Import Rows
(ingestFmv1ImportRows), then follow the job with Get Import Job
(getFmv1ImportJob) and List Import Attempts (listFmv1ImportAttempts)
filtered by jobId. To fix failed rows or change the mapping, queue another
job; for one record with an immediate result, use Import Entity
(createFmv1Import).
Queues transformations over stored import rows. The response contains the job ID and returns before mapping or writing. A new job is also how to retry or change a mapping; no re-import is needed. Jobs run in queue order within a company, with one attempt per result. Different jobs can map the same row into different entity types.
Expressions read row (the first source JSON value) and rows (all source JSON values for this result). Mapping starts with an empty target entity. Use groupBy to combine related rows. Within a result rows are ordered by storage time, then row ID; do not assume a source record kind is first.
Selectors are snapshotted when the job is accepted. Mapping and guard failures are recorded on attempts and processing continues. Queue dispatch failures leave the job queued for automatic recovery. GET the job and list attempts with jobId to follow progress.
Policy jobs require policyMode: whole-policy or single-transaction and refuse groupBy. Their transformations produce the corresponding policy command envelope. Both modes support execution and dryRun: true. Single commands return one receipt; whole-policy rows return ordered transaction receipts. See Policy import contracts.
Required permission: the target entity type’s import permission (policy.import for Policy, exposure.import for Exposure, for example). Selecting rows stored by the synchronous endpoint additionally requires their original entity type’s import permission. Permissions and API-key revocation are checked again by the worker.
curl --request POST \
--url https://go.aiinsurance.io/api/v1/companies/{companyId}/imports/jobs \
--header 'Authorization: Bearer <token>' \
--header 'Content-Type: application/json' \
--data @- <<EOF
{
"entityType": "Exposure",
"source": {
"importBatch": "october-load"
},
"transformations": {
"exposureName": "row.name",
"exposureType": "'company'"
},
"guards": [
{
"condition": "exposureName != null",
"errorMessage": "A name is required"
}
],
"attemptBatch": "mapping-v2"
}
EOF{
"id": "550e8400-e29b-41d4-a716-446655440003"
}Authorizations
User-principal OAuth 2.0 Bearer authentication. Send a user-scoped Auth0 access token (audience = the app API audience) as Authorization: Bearer <jwt>. The request resolves to the user's identity and is authorized by their Role on the {companyId} in the path — the same role-based permissions the web app enforces. This is the path the MCP connector uses to act on a user's behalf; endpoints that accept it list both BearerAuth and ApiKeyAuth.
Body
- Option 1
- Option 2
Quote jobs only, without groupBy. Restore an already bound legacy quote from its retained source identity, original bound/applied_to_policy status and existing policy. Existing quotes retain their identity and parent links; conflicting statuses or source evidence are refused. This does not bind a policy or generate invoices, recommended forms or lifecycle packets. See the historical bound quote contract in the import overview.
restore-bound How data.<entityKey> becomes an entity in this company's field model: one
entry per field reference id (camelCase). Each value is either a JEXL
expression string that reads the data's top-level keys by name
(submission.name, broker.email), or a mapping for a list of child
objects:
{ "mapOver": "<data path>", "values": { "<childKey>": "<expression>" } }.
Inside values, bare names and me are the current element, parent is
the object holding the list and root is the whole data. A child value may
itself be a mapping; its mapOver path reads the current scope (for example
me.locations). A { "values": { ... } } constructor without mapOver builds
one object in the surrounding scope and can contain mapped lists. At every
depth, root remains the original data.
Every field of data.<entityKey> without a transformation is kept as sent.
A transformed entity with an id updates that entity; otherwise one is
created. Keys the company's field model does not declare are dropped before
an entity write. Policy jobs instead construct the whole-policy or single-transaction
command envelope; envelope keys are validated by that contract.
Show child attributes
Show child attributes