The sound of the letterbox clatters. Brown envelope with a window, addressed visible through, no stamp. It’s the kind of mail that makes the room feel smaller before you’ve even torn open.
Aisha applied for Personal Independence Payment (PIP). The application was time consuming, requiring a detailed account of medical history that she knew was inside GP, Specialist Consultant and other Health and Care worker notes that Aisha had to contact, one by one and request her own historic records from. Some were emailed, most were posted on paper which she then had to photograph with her phone so they could be uploaded.
“Why can’t they just talk to each other?” She wondered.
The amount of questions were exhausting. How far she can walk, whether she can prepare food, whether she can wash, dress, manage medication, plan a journey, speak to people or cope with unfamiliar places, for someone with a brain condition she gets lost in every question and paragraph. Aisha has to call on family members to come round and help her fill in the form that she has printed off at the local Library because she cannot concentrate enough to do this on her phone. To apply, Aisha had to turn private difficulty into administrative evidence making pain legible to a test-based system. She had to describe fatigue without sounding vague, anxiety without sounding dramatic, and limitation without sounding incapable of everything. It’s an exam, one that adds to the already existing feeling of frustration of not being understood because her condition is invisible.
By the time the letter arrived, the service had already caused unintended consequences. She feels guilty of something, that having an impairment and asking for help is dirty.
From the perspective of the service provider, the process had been successful. Information was gathered, the evidence was assessed, a decision was made based on the rules. In the back-office system, the data was uploaded, a record was created, a decision made. The record in the system’s state changes, the system moves onto the next application. But Aisha is still where she was 3 months ago when she began looking for help to improve her quality of life.
Aisha doesn’t get extensive detail back to explain the decision, certainly not comparable to the amount of effort required to apply. Now she has to make a decision. Does she have the energy ,and the knowledge to challenge the decision? It will take time. Maybe she can try getting another job, or continue to rely on family members to help her remain independent.
A more detailed explanation may have helped. An easier application process may have meant it had felt less like an exam she had failed. A human conversation most certainly would have meant Aisha felt understood and guided whilst translating individual, unique circumstances into something that aligned with the rigid system. Instead, now she feels alone, ignored, and a burden on her family, her friends, on society.
Giving services relational context
The better service experience is not the same transaction made easier - even though connectivity between systems so that the burden of evidence is not on the individual would help. What could make a radical difference would be a service that helps Aisha remain oriented after the decision has been made. It explains what has happened in language she can use, connects the decision to what may happen next, and makes clear where responsibility now sits. It helps her understand what the service has recognised, what remains open, what support exists if the decision feels wrong, and what she should not have to explain again.
This is the alternative I want to explore now. After five articles arguing that transactional logic is too small for many public services, I want to turn towards what a better model feels like in practice.
For me, that begins with returning service design to its root idea - being of service to the person whose life is affected by what the system does next, whilst recognising the needs of the organisation so that it can be of-service.
That is why stewardship matters. It gives us a way to describe services that do not stop caring about consequences when the interaction is complete.
What good already looks like
Stewardship is not a new theory waiting to be invented. It already exists in fragments wherever services recognise that the meaningful work often begins after the formal transaction.
A diagnosis is not the end of cancer care. A referral is not the end of navigating a health system. The birth of a child is not simply a registration event. A financial product is not successful because the terms were disclosed. A care visit is not only a completed task. In each case, the organisation can either treat the moment as an output, or it can treat it as part of a continuing responsibility.
That is what I mean by stewardship: designing services so responsibility continues after many interactions. From here, I want to be more careful with the word transaction. It suggests a mutual exchange of value, yet many public services do not feel mutual to the person using them. They feel like obligations, assessments or judgements. For relational services, “interaction” is often the more honest word.
The word matters because it pushes beyond the language of delivery. Delivery asks whether something has been built, launched or improved - finished. Stewardship asks whether the service continues to hold responsibility for what happens next.
Service patterns in care pathways
Cancer care offers one of the clearest examples of this different standard.
Macmillan’s Holistic Needs Assessment is designed to identify and address the needs and concerns of people living with cancer, then develop a personalised care and support plan. In the patient-facing version, the person first answers questions or fills out an assessment, then talks through their needs and concerns with someone from the healthcare team, before agreeing a plan for care and support.
The design move is subtle but important. The assessment is not treated as a bureaucratic end point. It becomes the beginning of a conversation. The conversation becomes a plan. The plan becomes a way of carrying responsibility forward.
This is a useful pattern for public services because so many assessments behave as gates. They decide whether someone qualifies, what category they fit into, what level of support they receive, or whether the organisation can close the case. A stewarded service asks more of assessment. It still needs to be decided. It still needs evidence. It still needs fairness and consistency. But it also asks what the service has learned about what this person needs next.
For Aisha, that shift would matter. The decision would not tell her what she has or has not been awarded. It would help her understand what the service has recognised, what remains uncertain, what options are open, what support may be useful, and what should happen if her circumstances change. It would not turn assessment into therapy, or pretend that eligibility rules have disappeared. It would treat the decision as a moment of orientation rather than a hard administrative edge - she would know where she stands in a complex, ambiguous landscape.
In care pathways, good personalisation is not a decorative layer. It is the point where assessment becomes shared responsibility for what happens next.
Navigation is service at work
A second example comes from patient navigation.
Patient navigation developed because health and care systems can be difficult to move through, especially when people face practical, social, financial or emotional barriers. The World Health Organization describes patient navigation as an evidence-based personalised intervention designed to guide patients through complex cancer care systems so they can receive timely care.
I’m not suggesting that every public service needs a new role called a Navigator. The point is that navigation is recognised as part of the service. The person is not left to connect the systems by themselves.
This is where many public services quietly fail. We’ve spent a decade designing individual parts of a manufacturing production line without looking further out. We focus on building the application route, the guidance, the decision and the appeal process. Each part may be documented. Each part may be available. Each part may be technically accessible. But the work of knowing how to move between them still sits with the person, and a lot of the time there’s nobody in the organisation that actually knows how the entire system works.
In relational services, that connectedness, that human in the loop - that work is the service - knowing what to do next. Knowing who to contact. Knowing whether silence is normal. Knowing whether another service is affected. Knowing whether a delay is administrative or dangerous. Knowing whether to wait, chase, complain, appeal or start again. And critically that the person in need of support should never be the one having to do any of those things. Think of the number of times in your life where there’s been that one person who’s been around long enough that they become the oracle, the go-to person when something unexpected happens. In a service-led organisation that person should be everyone.
In navigation, the work is not to produce a smoother journey map, blueprint or clickable prototype. It is to design the connective tissue people are otherwise expected to supply themselves, or as Richard Pope describes it in Platformland, designing the seams.
Relational services are for life
Our final example of what ‘good’ looks like for relational services comes comes from Estonia’s work on proactive government services.
Estonia has been developing life event-based public services around moments such as marriage, divorce, the birth of a child, unemployment and retirement. The OECD Observatory of Public Sector Innovation describes this as a citizen-centric approach that reimagines how government agencies interact with citizens, organised around life events rather than agency boundaries. The principle is that people should not have to apply for services they are entitled to; where possible, those services should be offered proactively.
Estonia’s own digital government material describes proactive services as cross-authority event services, where several services related to the same life event are compiled into a single service for the user.
This is stewardship at the level of service architecture. The state does not wait for the person to understand its structure. It organises around the moment in life.
That matters because one of the burdens of public services is not only doing the task. It is knowing that the task exists. People have to know what they are entitled to, which organisation owns which part, whether one decision triggers another, and whether they have to apply, report, update, renew, challenge or wait.
For public service transformation, this may be one of the most important lessons. Better digital services are not only better individual transactions. Better digital services are better arrangements between organisations delivering a holistic outcome.
For Aisha, that does not mean pretending a PIP decision can automatically solve everything around her. It means recognising that the decision is not isolated. It may affect transport, care, housing, work, debt, advice, emotional wellbeing and other forms of support. A relational service should not behave as if its own boundary is the person’s boundary.
At the level of the state, good means understanding life events and changing circumstances well enough to bring support forward, rather than leaving people to discover every dependency by accident.
Patterns for more responsible services
These examples come from different sectors, but they point to the same design pattern.
Cancer care turns assessment into a plan. Patient navigation treats movement through the system as service work. Proactive government services organise around life events rather than institutional boundaries.
Together, they show that stewardship is not an abstract value. It is a quality a service can be designed to have - and that means it can be measured.
A stewarded service helps people stay oriented as their situation changes. It carries context forward rather than forcing people to begin again. It reduces the burden of joining things up. It recognises vulnerability without making people perform it endlessly. It treats waiting, challenge, appeal, complaint, renewal and change of circumstance as part of the service, not as awkward exceptions to it. Most of all, it asks what the person is left holding after the organisation has completed its step.
This is the alternative to transactional transformation.
For me, the practical test of a relational service is whether it leaves people more able to act.
That does not mean everyone gets the outcome they want. Public services make difficult decisions. But it does change the experience of being inside them.
A good relational service helps people understand where they stand, what has changed, what remains possible and what happens next. It helps them move through complexity without carrying every dependency alone. It reduces avoidable uncertainty. It treats silence as part of the experience. It notices whether people are becoming more confident or more lost as the service continues.
That is what Aisha needs after her PIP decision. She does not need warmth as a substitute for clarity, or to believe that the team has empathy, or that the development of the application form was tested with 100 people. She does not need a beautifully written letter that still leaves her alone with the consequences. She needs the service to hold enough responsibility for her to remain oriented, informed and able to act.
That is a better public service experience.
The future of service design
The last era of digital public service raised the floor, addressing a decade long lag in expectation from the online explosion. It challenged poor usability, unnecessary bureaucracy and institutional indifference to people’s experience. It made many services clearer, faster and more accessible.
The next era has to raise the ceiling.
It has to design for continuity as seriously as completion. It has to stop thinking about the journey as broken up parts so they can be reconfigured, redesigned and rebuilt to an Agile engineering process. It needs to stop thinking about triage, gates and handovers and behave more like water: Finding routes through complexity, carrying context with it, smoothing the hard edges that people currently have to navigate alone.
The design of services must remove itself from the manufacturing trap it has fallen into over the last decade, and get back to its roots being of service.
That is what stewardship means.