The Big Table Group

Challenge

When every hour is billed, change is slow.

On a time and materials model, every hour is billed. Each piece of work needs scoping, estimating, sign-off and invoicing — a cycle that repeats for every request and creates natural gaps where technical debt accumulates. Issues raised could take months to reach resolution.

For The Big Table Group, that meant meaningful platform improvements were very slow, and the pace of change struggled to keep up with where the business wanted to go.

The problem was never the engineering. It was a commercial model that made every improvement a negotiation.

Our Approach

A subscription for continuous improvement, with AI trained on their own platform.

We moved TBT onto XMS AI: one fixed monthly fee covering continuous development, with the same senior engineers in the platform every month. The engineering judgment didn't change. What changed was how much of it could be spent building rather than administering.

Generic AI tools start from zero, with no knowledge of a platform's history or architecture. XMS AI is trained on TBT's own codebase — their integrations, their infrastructure, and the reasoning behind past decisions. That context means our engineers start accurately from day one, and the model's precision improves as its knowledge of the platform deepens.

For one fixed monthly fee, all development, fixes, improvements and delivery happen continuously. No rate cards, minimal scoping overhead, no invoicing cycles.

The Outcome

More than four times the output, for a predictable cost.

Unlike a one-off cost saving, platform investment compounds. Rather than asking how to reduce the budget, TBT can now ask what they can build that they previously couldn't afford to.