Why service baselining is the missing first step for AI in Scotland’s public services
Scottish public services are being asked to do more with less, and the gap is widening. Audit Scotland’s most recent local government finance overview sets out the shape of the pressure: rising demand in social care, housing, mental health and additional support needs, against real-terms budget contraction and a workforce under strain.
The Scottish Government's Public Service Reform agenda makes clear that productivity, reduced duplication and service redesign will define the next decade of delivery. AI is increasingly cited as the enabler that closes the gap.
It rarely does, at least not yet. Most public sector AI pilots cannot demonstrate value, and the reason is structural rather than technical. Pilots tend to be evaluated on whether the model worked, whether the chatbot answered, whether the document was summarised. They are not evaluated on whether the service got better, because the service was never measured in the first place. Organisations end up with portfolios of technically successful pilots that have not moved any of the numbers that matter to citizens, to elected members or finance departments.
That is the missing first step.
Why traditional budgeting cannot measure AI value
Public sector budgeting reports on departments, teams and programmes. It tells you what the service costs in total. It does not tell you what it costs to process a single application, how that cost varies by application type, where in the workflow the time goes, or how much of that effort is spent on activities a machine could credibly support. Without that view, AI investment decisions are made against assumptions rather than evidence, and the business case for scaling what works is almost impossible to write with conviction.
A service-oriented view changes the unit of analysis. Instead of asking what the parking team costs, you ask what it costs to issue a blue badge. Instead of asking what the revenues team costs, you ask what it costs to process a council tax reduction claim end to end. Each of those services has a cost, a volume, an effort profile, a citizen waiting time, and all going well, an outcome. Attach data to those dimensions and you have the evidence base AI investment decisions need. You also have, not coincidentally, the evidence base that public service reform has required since the Christie Commission first set out its principles more than fifteen years ago.
The services where AI value is more likely to land
In our white paper, Automate Tasks, Not Jobs - The AI Opportunity for Scotland's Public Services, we identified the 50 services that underpin most local and national government delivery and rank highest for AI augmentation. Application processing, case management, enquiry handling, eligibility assessment, correspondence generation, complaints triage. These are the activities that consume the most caseworker time, generate the longest citizen waits, and recur across almost every public sector body in Scotland.
They share a common pattern. They take structured inputs, apply repeatable judgement, and produce written or transactional outputs at volume. That is precisely the territory where current generation AI is most credibly deployable, and where the evidence base for value is strongest. It is also where the case for shared learning across Scottish public bodies is most compelling, because the same service is being delivered in council areas, health boards and dozens of our national agencies, with broadly the same shape and broadly the same pain points.
Baselining is a leadership exercise
Service baselining is often described as an analytical exercise, but the part that matters is a leadership one. Baselining surfaces where effort goes, rather than where leaders assume it goes. It quantifies how much senior officer time is consumed by activities that could be handled differently, and how much citizen waiting time is generated by handoffs that should not need to exist. Those conversations are uncomfortable, but they are the conversations that turn an AI strategy from a list of pilots into a credible roadmap. They are also the conversations that give scrutiny committees, audit teams, and political sponsors the assurance they need that AI investment is grounded in evidence.
Where Storm ID fits in
Service baselining sits at the core of our AI Strategy and Consultancy work. We support senior leadership teams across local and national government in Scotland to map services, baseline current cost and effort, identify credible opportunities for automation and augmentation, and build AI roadmaps anchored in Public Service Reform priorities.
If you are responsible for AI strategy, digital transformation, or efficiency in a Scottish public body, we offer a free 90-minute baselining workshop with your senior team. You leave with an initial view of which of your services carry the highest AI value potential, an honest read on the evidence gaps you would need to close, and a sense of what a credible roadmap to capture that value looks like.
The prize
The opportunity in front of Scotland's public sector is a chance to deliver, finally, on the preventative, person-centred, joined-up services that Christie set out years ago, with tools that did not exist when those principles were written.
Service baselining is what turns that opportunity from rhetoric into a plan. Organisations that start with the baseline will know where their effort goes, where AI can move the numbers that matter, and how to evidence the case to scale. Organisations that skip it will end up with a portfolio of pilots and not much else to show for them. The difference between those two outcomes is the work that happens before the first model is deployed.