How our scores work
Every policy gets three scores, each from 0–100 and each shown with a plain-language band. Higher always means friendlier to you. All three are computed deterministically from stored data — never by a black box — so they're reproducible and auditable.
Clause Friendliness
· This policy's wordingRead purely from the policy document: how friendly the clauses, exclusions, waiting periods and sub-limits are to you. Computed deterministically from the AI-classified clauses — the same wording always scores the same.
Insurer Conduct
· This insurer's complaint recordThe insurer's overall complaint standing, derived from public IRDAI and Insurance Ombudsman (CIO) data — complaint rates, settlement delays, and ombudsman outcomes, normalised against peers. It is a company-level record across every line the insurer writes: for insurers that offer both health and general insurance (motor, home, fire, and so on), the same figure reflects all of that business, not health alone. It is a reputation signal about the insurer, not a verdict on this specific product.
Overall Policy Friendliness
· The headlineBlends the wording with the insurer's track record. A weak claims record pulls the headline down, but a good reputation can never rehabilitate genuinely predatory wording — and when there isn't enough recent conduct data, the Overall simply reflects the wording alone.
Bands lead, numbers support
The number is always shown with a band, and the colour comes from the band — so a 72 and a 74 read the same. A small numeric gap is noise; the band is the signal.
Comparable only within a type
Scores are calibrated per insurance type and the conduct signal is normalised within each segment, so a health score and a motor score are not on the same ruler. We only ever compare policies of the same type, and we never rank across them.
Where the conduct data comes from
Insurer conduct is built from public regulatory reports — the IRDAI Annual Report & Handbook and the Council for Insurance Ombudsmen (CIO) Annual Report. These publish about once a year, so every conduct figure carries the financial year it reflects and lags roughly 12–18 months. Each stored figure keeps its source links and a checksum so any number is traceable to the exact document, and every snapshot passes a human review before it is ever shown.
Informational only — not advice. KnowYourPolicy.ai's Insurer Conduct Score and Policy Friendliness Score are AI/automation-assisted interpretations of publicly available regulatory data published by the Insurance Regulatory and Development Authority of India (IRDAI) and the Council for Insurance Ombudsmen (CIO). The figures are extracted from public annual reports using automated OCR and statistical modelling, may contain extraction or interpretation errors, and are provided for general information only. They are not legal, financial, insurance, tax, or investment advice, not a recommendation to buy, avoid, switch, or cancel any policy, and not a statement of fact about any insurer's conduct or any prediction of how a claim will be decided. Source data reflects a specific past financial year and may be out of date. We are not affiliated with, endorsed by, or acting on behalf of IRDAI, the CIO, or any insurer. Always verify against the original regulatory reports and consult a licensed advisor and your insurer before making any decision. Use of these scores is at your own risk.