Previously, I outlined some challenges and an opportunity for service blueprints in the age of AI. I have also published a prototype, with a healthcare focus, to explore the challenge of evidence, or proof, in how we use AI in service blueprints.

In this post, I want to discuss a vision that I think is emerging in my network and explore its implications for work and for how we govern AI.

In short, the vision is a service blueprint that an organization uses to oversee the operation and change of its services. It ensures those services are aligned to human values and that the organization’s use of AI is governed by outcomes for the people it exists to serve.

The implication I want to discuss is what this means for human work. I hypothesize that it will require us to work in ways that are focused on value: empathizing, collaborating, problem-solving, and systems thinking. We will need to spend a lot more time together working out what is valuable, what is a priority, and what is working, and less time in the technical weeds.

Critically, this work on value, priority, and what works (and on judging what “works” means) is exactly what we need to do more of if we are going to govern AI. The real vision here is people doing this work and structuring it as a service blueprint: a shared, tangible record of the value, priorities, and work.

The vision

The service blueprint will become not just a design document but the reference design, governance mechanism, and, to use more AI-native language, system prompt for how organizations run their services. And if the blueprint is the system prompt, then whoever shapes the blueprint is, in effect, governing the AI. This is why the human work of defining value, priorities, and what works matters so much.

This is the idea I have discussed with colleagues at Livework and with Robb Wilson, founder of AI firm OneReach (coming soon on his Invisible Machines podcast). It is also a common sentiment in responses to my posts.

Imagine the organization of your choice. I’m choosing a healthcare system. In this future state, it has a service blueprint that is fully loaded (Robb would say “hydrated”) with all the data available about its patients, its people, its clinical activities, its technical systems, and external factors that may affect healthcare, such as the weather.

This data feeds a service blueprint that, at the highest level, is structured around the flow of patients. It starts with knowledge of the whole community and oversees how people stay well, access care, are cared for, manage change, and return home. This is the structure I have used for my healthcare prototype.

I get a sense of this through my experiments and in comments and discussion. In my prototype, I can synthesise a wide range of health outcome metrics from many sources with relatively little effort. But that ease raises harder questions. Which of these outcomes matter most? How should they be weighed against each other? And what would it mean for the service to be working? The easier it becomes to collate and structure data into a service blueprint, the more I feel the need for human engagement in defining value, priorities, and how things should work.

The implications

In this vision, what are people in the healthcare organization doing? Many are caring for patients and doing more of it because their time has been freed from administrative work. In general, the promise is that automation frees people for valuable work.

But AI should not be running a health system. Decisions about how to care for people, whom to care for, and what the best form of care is are value-based, and they belong with people. I hypothesize that this is the work people should be doing much more of, not only as carers on the front line.

So healthcare staff, from nurses to leaders, should in my view all be moving towards design and governance roles, facilitated by the service blueprint.

This is a big shift, and I don’t think it happens by default. Automation will free up time, but it won’t decide what that time is for. If we want AI to serve the people our organizations exist for, we need to invest in the human work of deciding what matters and give it a shared home in the service blueprint.

The article originally appeared on LinkedIn.