A product built from nothing but a problem.

The situation

Class action settlements are public. The problem is that being public is not the same as being findable. Notices are scattered across court records, administrator sites and legal notices written for lawyers, and claim windows close whether or not anyone eligible noticed them. The result is money set aside for consumers that quietly reverts because nobody filed.

There was no existing system to modernise here, no ERP to integrate with and no process to map. There was a hypothesis about consumer behaviour and a body of messy public data — which is a different kind of engineering problem, and a riskier one, because the wrong build is not obviously wrong until people fail to use it.

What we did

The technical core is unglamorous and it is where most of the value sits: continuously gathering settlement information from public sources, normalising records that were never designed to be compared, and deciding which of them plausibly applies to a given person. Getting that wrong in either direction breaks the product — miss real cases and it is useless, surface irrelevant ones and it becomes noise people stop opening.

The second half of the problem is the filing itself. Finding a settlement you qualify for is worth nothing if the claim form then defeats you. So the app keeps a profile, fills the official form from it, shows the completed PDF for review, captures a signature and tracks what happens next — turning an afternoon of admin into about a minute.

• Automated collection from public settlement sources, running continuously rather than as a one-off import
• A normalisation layer turning inconsistent records into something comparable and searchable
• Eligibility matching against a saved profile, so someone sees settlements that plausibly apply to them rather than a directory
• Auto-filled claim forms — the official form, populated from the profile, with a live PDF preview before anything is submitted
• Electronic signature, deadline alerts and claim status tracking, so a filed claim does not disappear into silence
• A cross-platform build shipping to iOS and Android from one codebase
• Subscription billing through a managed provider rather than built from scratch
• Plain-language summaries, because a settlement notice written for attorneys does not tell a consumer whether to act

Choosing a single cross-platform codebase and a managed backend was deliberate. For a product still proving its hypothesis, spending the budget twice on two native applications — or on infrastructure nobody had yet earned — would have bought very little.

Outcome

The result is a live consumer product on both mobile platforms with a continuously updated case set behind it — taking something that previously required knowing what to search for, and turning it into something that arrives. Most settlements it surfaces need no proof of purchase, and some pay out several hundred dollars.

Why it is here

Most of our work involves systems that already exist. This one did not, and it is on this page for that reason — the discipline of building something from a hypothesis is different from the discipline of replacing something that already runs, and both come up.