Our API delivery process is designed to accommodate all types of APIs. We work with central API services that connect to a backend, intermediary APIs that route requests or messages from a sender to a receiver and APIs that are documentation only (with no central backend). Part of my job is to work with the API producers to decide what is applicable and will add value to their program.   

Our ‘process checklist’ is at the core of the API delivery process.  It’s intended to be a self-serve resource supplemented with information from our ‘Producer Zone’ in Confluence (the documentation tool we use). It has a series of steps designed to guide the delivery teams through the API delivery phases. These aren’t prescriptive, just recommended steps. We work together to decide whether these will add value to a producer’s API.

By having a repeatable process, we cut the time and cost of building APIs. A consistent process means we can get our producers onto the platform quickly and get their consumers up and running fast.  

We have light-touch governance checkpoints. Teams wanting to use the API platform have a kick-off meeting with us, where we go through admin, explain the API delivery process, introduce the checklist and the Confluence resources.          

In discovery, teams engage the wider API management team in an architecture review. Our architects have a ‘birds-eye’ view of APIs on the platform and can highlight opportunities and possible tripwires.  

In ‘alpha’, as the producer starts to build their API, there’s a design and specification content review with the API management team. This aims to maintain consistency across the platform.  

Finally, in ‘beta’ we help producers through a ‘first time production review’ to try to smooth the journey to getting the API live.  

We have exit reviews in the discovery, alpha and beta phases, which are as simple as agreeing what steps are applicable and will add value. Often, some of the phases will run in parallel, so completion of a review is never a blocker for delivery. The key is always to make that journey smoother for our API producers. 

We provide technical support through our dedicated producer Slack channel, so producers have always got someone to talk to, and I have regular 15-minute catch-ups with producer teams to help them through the process and remove obstacles.  

Toward the end of the API delivery process, we run sessions with the API producers to capture their experience.  

They particularly like the support we give them on Slack and the short, focused catch-ups. They have also praised the guidance provided by our ‘process checklist’ and our extensive documentation.  

Up to now, we have mainly worked with internal producer squads within API Management and teams in other parts of NHS Digital. The API platform is now available to API producer teams outside NHS Digital if the use case is appropriate. The service is going global. There have been recent discussions with the Ministry for Health in New Zealand, who are keen to learn from our experiences. 

If you have anything to say to us, either as a producer or a consumer of our APIs, get in touch with myself or Tony Heap at [email protected]. We are particularly keen to learn about ways we can improve.