The short answer: if your brokerage or MGA still depends on a legacy policy admin system, 2026 is the most affordable moment yet to modernise. AI-assisted migration tooling now does much of the heavy lifting - analysing old applications, extracting business rules and mapping processes - and analysts estimate generative AI has cut the cost of building production systems by roughly 90% compared with 2022. The barrier that made "we'll migrate someday" rational has largely gone.
What's the state of legacy software in insurance?
The numbers are striking for an industry built on data:
- Over 70% of insurers still operate on legacy platforms
- 54% of insurance IT budgets go to maintaining existing systems rather than building new capability
- Most core insurance suites run on technology that was modern in the late 1990s
Boston Consulting Group's 2026 research describes leaders now treating core modernisation and operating-model change as one synchronised programme - with agentic AI doing discovery, rule extraction and documentation that used to consume months of consultant time.
Why does this matter for a small broker, not just the giants?
Because the pain is proportionally worse at small scale. A 10-person brokerage doesn't have an IT department to nurse an ageing system - it has one "person who knows how it works" and a growing pile of workarounds: re-keying client data between the policy system and spreadsheets, manual renewals chasing, documents scattered across shared drives. Each is a quiet, recurring cost, and each is now automatable.
We've migrated businesses off specialised legacy platforms - the same pattern applies whether the system is Proclaim, Act or a bespoke tool nobody dares touch. The details are in our note on legacy software and human transformation.
Do you have to replace everything at once?
No - and for an SME you almost never should. The lower-risk path we recommend:
- Integrate first: keep the legacy core, automate the data flowing in and out of it (renewals, documents, client comms)
- Digitise documents: scanned policies and correspondence become searchable, structured data - see our guide to automating text recognition
- Migrate last: once the processes around the core are automated and understood, moving the core itself is a far smaller, better-specified project
This sequence means every stage pays for itself before the next begins.
What does modernisation cost in 2026?
Far less than the quotes you may have filed away from a few years ago. AI-assisted analysis compresses discovery, and cloud infrastructure removes the upfront hardware. For a typical SME broker the first automation phase is a weeks-long project, not a year-long programme - and the maintenance line it replaces (54% of IT spend, industry-wide) starts shrinking immediately.
FAQ
Is it safe to migrate client data off an old system? Yes, with a staged approach: data is extracted, validated and run in parallel before anything is switched off.
What if our supplier no longer supports the software? That's a reason to start sooner - unsupported systems carry compliance and security risk that grows every year.
Do we need to be technical to manage this? No. The point of a well-run modernisation is that your team keeps working the way they like, while the re-keying and chasing disappears.
The easiest way to start: pick the workflow that annoys your team most, and ask us for a proof of concept. We'll automate that one process against your real system, so you can judge the benefit before committing to anything larger.