Podium | Our Experience

Podium's data feeds are used by tier 1 bookmakers in the UK but their technology infrastructure hadn't been touched in 20 years. A modernisation their own leadership estimated would take three years was implemented by CreateFuture in 7 months at half the cost.

Meet Podium

Podium supplies the sports and racing data that sits behind live betting markets across the UK. Its feeds price and settle races for every major UK bookmaker, and more than 300 operators, media companies and racecourses use its content and technology day to day.

Horse racing is where the business built its reputation. Podium holds over a century of historical racing data, a depth that's hard for anyone else in the market to match. It's also, as this case study shows, the reason a legacy system built up over twenty years became such a high-stakes thing to touch.

The challenge

For years, Podium had been producing the odds shown on every major UK betting platform, but the system was starting to creak. Output was running on four separate legacy systems, some written in C and Perl on ageing Solaris hardware. All this was held together by manual workarounds that only two people fully understood.

Operators were checking every incoming race record by hand, and there was no automated way to tell good data from bad. Removing that overhead was the one thing standing between Podium and scaling its commercial potential without needing to scale headcount alongside it.

Without AI-native acceleration, conventional full-platform development of this scale typically spans years, delaying value and Podium needed value sooner. Fixing this called for skills that rarely sit together in one team: real-time data pipeline engineering, schema normalisation at scale, and the patience to reverse-engineer systems that were never written down in the first place.

The approach

CreateFuture proposed Project Nexus, a single API gateway replacing multiple disconnected integrations. A unified GraphQL and REST API, a self-service developer portal, and a sandbox environment meant new customers could integrate and test without Podium engineering support.

Proving that the new system would work was essential. It was tested end to end on Podium's highest-volume, most data-intensive category: horse racing.

What a traditional consultancy might spend four weeks establishing, CreateFuture achieved in four days. A six-person team, not the ten to fifteen a conventional build would carry, worked from Podium's offices alongside their operations, technology, product and commercial leads from day one.

AI-native delivery

AI was the infrastructure the delivery ran on from day one, changing the throughput, quality and knowledge retention achievable with a small team.

Before any production code was written, the team set up a version-controlled knowledge base: every architecture choice, data-quality rule and delivery decision recorded and queryable, with every AI agent session reading from it and writing back to it. This allowed AI tooling to read Podium's undocumented legacy code directly and reverse-engineer systems, such as an internal race-imagery tool with no formal documentation, rebuilding them in the new stack in weeks rather than months.

Automated feedback loops checked every agent-generated change against Podium's business rules and acceptance criteria before it reached a person. That gave agents a way to correct mistakes early, and meant engineers spent less time reviewing pull requests that were incomplete or already going to fail.

The result

CreateFuture delivered a production-grade platform that's AWS-native, covering the full horse racing category from data source to customer. We also built multiple sports integrations and a multi-sport data platform covering the full range of sporting events. The payoff is that ops teams get one familiar interface across events, which cuts training time and lets people work across event types instead of specialising in just one, so it's a real efficiency gain.

Operators now review only the 1% of incoming data flagged as needing attention. The other 99% is cleared automatically, eliminating the manual overhead that was directly suppressing margin. Data integrity is enforced at source: duplicate entries are resolved automatically, so a horse appearing under variant spellings is recognised and merged. For a business whose feeds underpin live betting markets, that's a reputational obligation.

New customers can now integrate through a single API and self-service developer portal instead of five separate systems, cutting a new data supplier onboarding from twelve weeks to two. Weekly sessions with Podium's engineers and a hackathon that shipped three live features in a day mean Podium's own team can extend the platform to new sports without relying on CreateFuture to do it for them.

Ongoing engagement

The same architecture is now being extended to football, tennis, Formula One, basketball and international racing. These projects are being run by a team that already knows how it works, because the architecture built for horse racing was designed to scale exactly this way. The results will be the same: margin improvement, operational savings, and commercial growth capacity.