What people actually want from an insurance system
Not features — four numbers.
Ask an insurance software vendor what they sell and you will hear about features: dashboards, workflows, portals, artificial intelligence. Ask the person who signs the contract what they bought and, if they are honest, the answer is different. A coverholder, an MGA, an underwriting agency — none of them runs features. They run a profit and loss statement. A system is worth exactly what it does to that statement, and nothing more.
This matters because most technology differentiation in insurance is decoration. A cleaner interface, a smarter document tool, a better-looking report — pleasant, occasionally time-saving, economically invisible. The buyer’s business looks the same at year end with it or without it. Real differentiation is narrower and harder: it moves a number that decides whether the business makes money. These four numbers explain most buying decisions we see.
The four numbers
Cost to start
Capital spent on systems before the first dollar of premium — licences, rating, documents, portals, bordereau tooling. Every dollar raises the GWP required to break even, and it is all spent when the risk of failure is highest.
Cost to run
One to two back-office people producing bordereaux and reconciling cash, tax and remittance in spreadsheets, plus the audit preparation a manual workflow makes expensive. This cost grows roughly in line with the book.
Cost to connect
Every trading platform, broker system or distribution partner is a bespoke integration project. At bespoke prices most partnerships are never attempted — the cost caps the firm’s reach.
Revenue per person
In flow business the fastest quote wins a disproportionate share. When quoting takes days, small premiums are uneconomic to write, and each underwriter’s capacity — and therefore revenue — is fixed by manual throughput.
There is a fifth item that behaves less like a number and more like a ceiling: capacity. For a delegated-authority business the binder is the revenue base. It is renewed, expanded or lost largely on the quality and timeliness of the data flowing back to the capacity provider. A firm can get everything else right and still watch its revenue ceiling drop because its bordereau is late, wrong or unallocated.
What that implies a system must actually do
Move cost off the starting line and onto the book
The cost of the system should travel with the business it carries, not sit as a large fixed base recovered before the first policy. That single change moves breakeven to a fraction of the GWP it sits at today. It also changes who carries the risk of a slow start — which is why our own fee basis is a subscription scaled to the book plus usage that follows the work, rather than a capital project up front.
Build once, share with everyone
Rating tables, tax and levy logic, bordereau formats, reconciliation, compliance gates and policy wordings are today rebuilt, at full cost, inside every firm in the market. There is no commercial logic in that duplication. Infrastructure built once and amortised across every customer delivers each firm a stack it could never justify building alone.
Make connection cheap
When an integration built for one counterparty is reusable by all, its cost falls by an order of magnitude. At that price, distribution that was commercially impossible becomes routine, and every new connection benefits every firm on the platform at once. Reach compounds; the cost does not.
Make speed the default
A submission that goes from proposal document to issued quote in minutes changes the revenue line twice: hit rate rises because the fastest quote wins, and whole segments of small-premium business that were uneconomic by hand become profitable at volume. The same underwriting team writes multiples of its previous count — growth without matching headcount.
Make compliance a by-product
If every bound risk is already classified, coded, split and reconciled the moment it binds, then the bordereau, the tax return and the audit trail are outputs of ordinary trading rather than projects bolted onto it. The back office shrinks, audit preparation drops from weeks to hours, and, crucially, the capacity provider gets clean data on time, which is what protects and grows the binder.
The test. For any claimed differentiator, ask three questions. Which of the four numbers does it move? By how much? And who keeps the money? If the honest answer is not a lower breakeven, fewer people per dollar of premium, cheaper reach, more premium per underwriter, or a safer binder — it is decoration, whatever the brochure says.
The conclusion
What people actually want from an insurance system is simple to state and hard to build: start without a capital project, run on fewer people, reach more distribution than they could afford alone, quote faster than their competitors, and keep their capacity provider satisfied without trying. Each item on that list is a hard economic outcome, not a feature.
That is the standard any system — including ours — should be judged against. Not what it does, but what it changes on the page that matters.
Cuttleflow Systems · Perspective 01 · 33°53′S · 151°16′E · Sydney