ClaimCompass · API & MCP
Biopsychosocial scoring your own systems can call.
ClaimCompass is the scoring engine behind an income-protection claims platform, taken out and offered on its own. It is built on the biopsychosocial model used across UK and reinsurer claims practice, and it returns a score with the reasons behind it. Call it from your claims system or data warehouse by API, or hand it to your own assistants as a set of tools.
What you send, what you get back
One call does the work.
Score a case
Send the questionnaire answers and the case facts. Get back a risk band (Low, Moderate, High, Very High, or Pending when the data is thin), four domain scores — biological, psychological, social, administrative — an intervention-need index, a confidence index that says how much of the answer rests on real data, the top drivers behind the score, and a prioritised list of next actions for the case manager. Stateless: nothing about the claimant is stored; one audit row records the call.
Score and render
The same call returned as a branded PDF for the file, with the personal-data whitelist applied.
Hand over a claim
Push a claim from your system under your tenant key. It comes back with an assessment state and the gaps named, so the missing information can be routed to the claimant or the case manager.
Where the method comes from
Evidence you can trace.
Your data, your decisions. The engine returns a score and the reasons; a person in your organisation decides. It is not a benefit determination and does not replace the case manager.
- Waddell's biopsychosocial model of work absence
- Validated instruments for work ability, mood, anxiety, self-efficacy, fear-avoidance and social support, cited item by item
- The 23-item psychosocial research framework published by Swiss Re in 2015, mapped to the questionnaire
- Every rule that changes a score names its source and its version, so a reviewer can see what moved and why
Two ways in
By API today, or as tools for your agents.
API
Bearer key per organisation, every call logged, rate-limited, contract versioned. Sandbox key on request.
MCP server
The same calls packaged as tools for your own AI agents, set up for your estate, with a person approving anything consequential.
For claims operations, rehabilitation providers and case-management firms that already have a system of record and want the capability inside it. Pricing on request, after we have seen what you would call and how often.
What ClaimCompass proved
The platform ran in internal beta with a small number of organisations: structured referrals in place of emailed forms, one auditable case across insurer, provider and claimant, session approval and outcome tracking. The scoring engine, the questionnaire and the audit trail are what moved into the service.
Start
Ask for the endpoint contract and a sandbox key.
Tell us what system would call it and roughly how often. We reply with the contract and a key for a sandbox tenant.