Patterns
The same patterns work in a sports club and in a large organisation; only the scale and the controls change. Each one is a lesson from our products and its twin in enterprise work.
Keep customers or environments apart
- In our products: many clubs or businesses share one system, and none can see another's data. A test checks every screen as a brand-new customer.
- In enterprise work: separate environments with no reads or writes across them.
Configuration, not code
- In our products: commission rates, feature switches and an emergency off switch are settings, not releases.
- In enterprise work: data pipelines driven by configuration, so a new feed needs no new code.
Let the business change things safely
- In our products: club and business owners run their own branding, features and members.
- In enterprise work: data stewards update reference data through reviewed changes, deployed with the code.
Code decides, AI writes
- In our products: AI assistants pick who and when in code; the model only drafts the words; people approve anything that can't be undone.
- In enterprise work: any AI use where decisions must be explained to an auditor or a customer.
One engine, many uses
- In our products: one shared engine runs 15 AI assistants, each with the same limits built in.
- In enterprise work: reusable frameworks instead of one-off pipelines.
Build with AI, review independently
- In our products: AI writes code under automated checks; a separate AI reviewer checks each change; nothing broken can deploy.
- In enterprise work: AI-assisted engineering a team can adopt without losing control.
Pay only when it runs
- In our products: assistants run as a scheduled job; AI spend is capped per customer; build minutes are rationed.
- In enterprise work: cloud cost discipline.
Move with nothing lost
- In our products: an old website plus separate booking and coaching systems folded into one platform.
- In enterprise work: moving off old systems in stages.