APIs & event-driven connections integrations

Use a supported interface to exchange the records needed for the workflow. Authentication, duplicate events, version changes and update limits are part of the design. Fast communication is useful only when both systems interpret the event correctly.

A stock change, payment outcome or dispatch milestone can trigger an update in another system. The connection should recognise the same event if it arrives twice and preserve the original reference so staff can trace the result.

Check authentication, available operations and the source system’s limits. A webhook can notify another service quickly, but processing delays and failed requests still need a queue, retry policy and visible exception route.

Choose representative records with the people who use the system. Include a variant, a changed value and an exception relevant to this workflow. Check the resulting record in both systems and confirm who can correct it.

Agree how the team will notice a delayed connection and what should happen until it recovers. Where updates are scheduled, show that limit in the operating plan rather than describing the result as instant.