REST API for Insight invites and automation
The REST API lets engineering trigger personalized Insight invites from product events. Keep research in the same loop as billing, support, and lifecycle tools — conversational interviews with blueprint capture and editable analysis, without building a custom survey stack or buying a panel recruitment API.
Owned-audience research works best when the invite arrives in the same week as the moment that matters: checkout complete, ticket resolved, day-seven no activation, usage drop before churn. The REST API is how your backend creates or matches a contact and sends a branded Insight link without waiting for a manual CRM export.
Every session still runs through the same Insight agent you configure in admin — objectives, session shape, blueprint fields, optional voice, and the analysis hub with editable auto-metrics. The API is distribution and automation, not a separate interview product.
Teams usually start with public links or CRM bulk invites on free sessions, then wire API triggers once themes prove useful. That order keeps engineering time behind evidence instead of ahead of a speculative integration project.
When to use this
- A product or billing event should open a research window within hours, not at the next quarterly wave.
- You already store plan, region, or risk flags in your system of record and want those fields on the contact for hub filters.
- CS or lifecycle tools fire webhooks you can turn into Insight invites without rebuilding interview UI.
- You need server-side secrets and scoped keys — never embedding credentials in the participant frontend.
Example use cases
Post-checkout or post-activation interviews
When an order pays or an account activates, create the contact and send a pulse or deep-dive Insight invite while the experience is fresh.
Churn-risk and usage-drop triggers
On integration disconnect or engagement drop, invite the mid-tier cohort into a churn-risk agent with blueprint fields for save willingness and competitor named.
Support ticket close follow-ups
After a ticket resolves, trigger a short CSAT-style Insight session so agents capture whether the fix worked — not only a star widget.
Setup steps
- 1
Create an API key
In company settings, generate an integration key with the scopes your backend needs. Store it as a secret in your vault or environment — never in frontend bundles, public repos, or embed snippets.
- 2
Publish an Insight study
Configure objectives, session shape (pulse vs deep dive), blueprint fields, branding, and optional voice. Note the survey or agent id your API calls will reference.
- 3
Map product events to research moments
Decide which backend events deserve an invite: order paid, ticket resolved, day-seven no-activation, usage drop, plan change. Keep the list short so completion stays high.
- 4
Call invite endpoints from your backend
On the event, create or match a CRM contact and send a personalized invite link. Prefer idempotent handling so retries do not spam the same person.
- 5
Sync contact metadata for analysis
Pass plan, region, cohort, or risk fields so analysis hub filters and compares match how you already segment customers in billing or CS tools.
- 6
Monitor completions and close the loop
Use admin KPIs, analysis digests, or status checks from your side to confirm sessions complete and notes are ready — then act on editable metrics in the hub.
FAQ
Do I need the API to run interviews?
No. Public links, CRM bulk invites, and website embed work without code. Use the API when events in your product should trigger research automatically.
Is the API a form-builder webhook only?
No. It supports Insight agent distribution for conversational interviews — blueprint capture and editable analysis included — not just dumping answers into a spreadsheet.
Where do I find docs?
API documentation lives with your admin workspace and help center. Start with integration keys, then invite and contact endpoints.
Can I start on free sessions before wiring the API?
Yes. Many teams prove themes with free Insight sessions and CRM invites first, then add API automation once the research loop is worth engineering time.