Risk intelligence for health and insurance
We build risk stratification, risk management and claims integrity tools for insurers, health payers and providers. They point people's attention to where it can do most good, and they leave consequential decisions with people.
Last updated
Draft (F11): Confirm the definitions reflect what we build.
What the terms mean
- Risk stratification
- Grouping people by likely future health needs, so care and outreach can be prioritised.
- Risk management
- Acting on those signals.
- Risk optimisation
- Directing resources where they do most good.
Claims integrity
Spotting patterns in claims that may indicate fraud, waste, over-prescription or overcharging, for a person to review.
Draft (D4): Averra's risk stratification with the Johns Hopkins ACG System, in the partner's approved wording.
In Averra
Averra uses the Johns Hopkins ACG System for risk stratification, under licence.
Safeguards
- Human review
- People review consequential decisions. Our tools inform them; they do not make them.
- Explanations
- Signals come with the reasons behind them, so a reviewer can check them.
- Fairness checks
- [TO CONFIRM: fairness checks we run]
- Privacy
- Sensitive data is anonymised or pseudonymised, and minimised, wherever the work allows.
Regulation
- Under the EU AI Act, AI used for risk assessment and pricing in life and health insurance is classed as high-risk (Annex III), with obligations currently due from 2 December 2027.
- UK GDPR sets rules on automated decision-making about people.
- UK insurers also work within the Financial Conduct Authority's rules, including the Consumer Duty.
We design with these in mind.