Policies keyed. Claims read. Portals worked — by an agent.
Broking runs on data moving between systems: policies keyed into the admin platform, claims documents read and filed, quotes pulled from insurer portals. It's accurate, repetitive work that ties up your team all day. Nova builds AI agents that do it — operating the software you already use, no API required.
Your team spends the day re-keying instead of broking.
Every new policy is typed into the admin system by hand. Claims arrive as documents someone has to read and file. Quotes mean logging into three insurer portals. None of it needs an account handler's judgement — it's volume, and it's the most expensive way to spend their time.
Exactly what changes.
New policy data entry
Before
Account handlers re-key every bound policy into the admin system by hand.
After
Agent logs in, keys each field, and confirms with a screenshot. Handlers check only exceptions.
Claims document processing
Before
Claim forms and documents read and filed manually, one at a time.
After
Document read → fields extracted → validated → filed in the system automatically.
Insurer portal quotes
Before
Staff log into multiple insurer portals to pull quotes for each risk.
After
Agent logs into each portal, pulls the quotes, and brings them back for comparison.
Renewal & MTA migration
Before
Renewals and mid-term adjustments re-entered between systems by hand.
After
Agent reads the source and re-keys each change accurately, record by record.
New policy → keyed into the admin system
When a policy is bound, the agent logs into your policy admin platform and enters every field the way a handler would — insured, cover, premium, dates — and saves a screenshot of each entry as an audit trail. Anything it isn't sure about is flagged, never guessed.
Claim → read → filed
Claim forms and supporting documents are read on arrival, the structured data extracted and validated against your rules, and the record raised in your system. Your handlers see only the exceptions, not every routine claim.
Risk → portals worked → quotes returned
For each risk, the agent logs into the insurer portals you use, runs the quote, and brings the results back so your broker compares options instead of spending the morning logging in and clicking through.
Renewal → migrated → confirmed
Renewals and mid-term adjustments are re-keyed between systems automatically, record by record — the kind of bulk migration that normally eats a week of someone's time.
Hours of policy and claims keying returned to your handlers, quotes pulled without a person logging in, and a clean audit trail on every entry — measured against your own throughput during the diagnostic. UK-based, with your data kept in the UK.
Questions we're asked
Does it need an API to our policy admin system?
No. The agent operates the system through the screen, the way a handler does — so it works with platforms that have no API, like Acturis or an in-house admin tool.
What happens when it's unsure about a field or a claim?
It flags the item and routes it to a person with full context. You set the confidence thresholds — ambiguous data is never pushed through.
Where does our client data go?
It stays in the UK. We run on UK/EU infrastructure and your data is never used to train any external model — which matters for FCA and data-protection obligations.
What does it cost?
It depends on volume and the systems involved, which the diagnostic scopes. Engagements are outcome-bound — we agree the result first. Submit an application and we'll scope it.
If that's the constraint worth removing, apply.
We review every application ourselves and respond within two working days. If there's a fit, we start with a diagnostic to confirm where the real bottleneck is.