We’re starting this month’s article with a story from Changing Futures Northumbria. The person has been anonymised by the original author, Mark A. Smith.
Mary lived alone, with very little contact from anyone. A family member was fraudulently claiming her benefits and sticking a small amount of the money through her letterbox. Her oven was broken, so she made everything with a slow cooker. Sometimes the food was undercooked and she was often ill.
Her neighbours saw her being sick in the back garden and raised concerns through a helpline, but the response was that Mary had to refer herself for help.
Mary did not know what self-referral meant. She was not calling someone with a clearly defined need, ready to select the right category, provide the evidence and complete the process. She was living inside a set of connected circumstances that were slowly becoming dangerous.
The person who finally changed what happened was a gas engineer carrying out a mandatory safety check. Concerned by what he found, he contacted the Liberated Method prototype operating in Gateshead. Someone visited Mary at home.
They did not begin with an assessment form or ask her to identify which service she needed. They began by understanding what mattered to her.
Mary wanted a working oven because she liked to bake.
From there, the team brought in help. Her GP became involved. The police investigated the missing benefit payments. The house was cleaned and furnished, the kitchen became usable and Mary began receiving the money to which she was entitled. As she regained stability, the team stepped back, but not away. Over time, she went out more, connected with neighbours and started baking again.
The service was not the gas check, the benefits investigation, the visit from the GP, the cleaning of the house or the repair of the kitchen. It was the relationship that allowed those things to become a coherent response to Mary’s life.
Electricity is energy, not technology
Last month, I argued that we have spent the last 25 years building better digital candles. We have taken existing bureaucratic processes, moved them onto screens and called the result transformation. The interface became cheaper and more convenient for the organisation, while people remained responsible for carrying information and intent between disconnected parts of the system. We made people the human API.
Generative AI risks continuing the same pattern. Tutorials promise to make Claude or ChatGPT work for you, but often return the burden to the person: write exhaustive instructions and keep correcting what comes back. The technology changes while responsibility for the mechanism remains with the user.
I ended by asking what the electricity of services might look like. Could it be an anticipatory system that recognises a life event and coordinates a response before someone has to ask?
For Mary, electricity did not look like an app, a shared database or an automated notification. It looked like a person who could understand the whole situation and remain involved while different parts of the system responded. For Mary, as for most of us, the service was human.
The GP, police, benefits system and housing support all had something necessary to contribute. The difference was in how their work was assembled. Mary did not have to understand the architecture, approach each service separately or repeatedly prove that she belonged within its scope. The team absorbed that complexity around her.
That is what the wiring in your house does. It creates the environment through which electricity can flow. The complexity moves into the infrastructure so that you can experience the outcome.
Understand, don’t assess
The approach used in Mary’s story is called the Liberated Method. It developed through prototypes in Gateshead and was later used across Changing Futures Northumbria to support people experiencing multiple disadvantage.
The name comes from an attempt to free frontline workers to use creativity, compassion and judgement. It is not permission to do whatever feels right; a small number of rules and principles shape the work.
One principle is to understand, not assess.
An assessment begins with the service: does someone meet its criteria, which category applies and what resource can be offered? Understanding begins with the person. What is happening in their life? What matters to them? What would a good life look like? Only then does the work explore which knowledge, relationships and resources might help.
I experienced this during the response to the Grenfell Tower disaster. In the months afterwards, I worked with caseworkers seconded from across the country and assigned to support individual families. In a rare model for today’s social work, a family had a dedicated caseworker who was there for them.
They found clothes and got children to schools that, for some, were now more than 20 miles away. They organised prepaid cards, made dinner, spoke to embassies, helped relatives travel to England and looked for homes so families did not remain split across hotel rooms for months or years.
No web form could have done that work. An AI chat agent could not have held that responsibility.
Assessment turns a person into a version the service can process. Understanding asks the service to respond to the person who is actually there.
It also changes what counts as useful information. A broken oven could be dismissed as outside the scope of health, benefits, policing or social care. In Mary’s life it connected nutrition, illness, poverty, confidence and something she valued doing. It was not a minor domestic detail. It was an opening through which a different future became possible.
The relationship sets the scope
Traditional public services define their scope in advance. Housing deals with housing. Health deals with health. The benefits system deals with entitlement. Each boundary may be reasonable from inside the organisation, but lives do not respect those boundaries.
When circumstances overlap, a referral is meant to connect the pieces. In practice, it often transfers responsibility. The person moves between professionals, repeats their story and is assessed again through a different institutional lens. They can become surrounded by specialists without anyone holding an understanding of the whole.
The Liberated Method reverses that movement. The caseworker holds the relationship and pulls specialist help into it rather than referring the person away. The expertise remains, but continuity and responsibility stay in one place.
This is a practical redistribution of power.
The worker has enough authority to respond to what they learn rather than being confined to a predetermined menu. The person can define what a better life means instead of being reduced to outcomes the organisation already measures. Specialists contribute without taking ownership of the whole person. Leaders carry the risk of enabling judgement rather than pushing it onto the worker or the public.
None of this works if the organisation only asks frontline staff to care more. Workers cannot maintain a relationship if caseloads, targets and commissioning rules force them to close, transfer or narrow it. They cannot use judgement if variation from process is treated as failure.
Good relational work depends on organisational design. We saw this in July through Buurtzorg’s neighbourhood nursing teams. Gateshead takes us further, to the boundary between organisations where one relationship holds the work together.
Stepping back, but not away
There is an obvious challenge. Does a continuing relationship create dependency? Does it replace one form of institutional control with another?
Changing Futures Northumbria describes the work as balancing external support with someone’s own agency and capacity to change their life. When a person is in crisis, workers may need to act extensively to create safety and stability. As the person becomes more able to act, the support changes.
Mary’s team stepped back, but not away. A transactional service tends to be either active or closed. A relational service can change its intensity while retaining continuity, recognising that progress is rarely a clean sequence and people should not have to begin again after a setback.
The outcome is not permanent support. It is enough stability, connection and confidence for people to shape their own lives. The relationship should increase agency, not absorb it.
Not every situation needs this response. The published cases include someone with strong family and community connections who only needed a few ordinary transactions. Relational must not become another universal model. The point is to build services capable of responding to variety.
Place is becoming part of the reform agenda
This example feels especially relevant now.
In July 2026, the government published guidance for five Place Based Budget pilots in England. The programme starts from a familiar diagnosis: public services have been organised around the departments funding them rather than the lives of the people using them.
Its principles are integration, prevention and devolution. Mayoral Strategic Authorities will bring local partners together, pool elements of funding and redesign support around shared challenges. One pilot focuses on young people at risk of offending in Gateshead and South Tyneside.
This is a promising change in the unit of design. Instead of asking how one department can optimise its service, the pilots ask how organisations in a place can share responsibility for an outcome.
But they remain tests. Pooled budgets, governance boards and shared outcomes will not automatically produce relational services. Organisations can coordinate their processes while leaving people to experience the same assessments, referrals and handovers underneath.
The deeper test is whether place changes where power sits. Can someone close to a person’s life act on what they understand? Can support cross a boundary without transferring responsibility? Can local leaders change rules that obstruct the right response? Can the system learn from frontline relationships rather than dictate how they function?
If not, place-based delivery risks becoming another layer in the architecture: a larger administrative boundary containing the same service factories.
Gateshead’s experience gives these pilots something tangible to learn from. Place-based reform cannot live only in budgets and governance. It has to reach the relationship between a person and the public institutions there to support them.
What changes on Monday?
System change can feel impossibly distant. Legislation, funding, professional boundaries, data-sharing agreements and accountability structures are difficult to move. That can turn the existing system into a convincing reason to postpone changing anything.
The Liberated Method challenges that excuse without pretending the structural barriers are imaginary.
Some changes require different commissioning, governance and funding. Others begin by examining the work already happening. Where is assessment being used when understanding would produce better information? Where does a referral transfer responsibility rather than bring in expertise? Where are staff bending a process to produce a humane outcome, and what does that tell us?
Service designers already observe these workarounds and try to incorporate them into what comes next. We call them hacks.
Our raw materials are not only interfaces, content and digital journeys. They include the scope and authority of a role, the rules guiding judgement, the way risk is shared, how cases are reviewed and whether leaders learn from what workers encounter.
A service blueprint can show a handoff. The more important work may be deciding that responsibility should not be handed off at all.
The infrastructure is a relationship
Mary’s story is not a model to copy unchanged. It emerged from a particular place, programme and group of people. The evidence is encouraging, but the wider evaluation also describes the difficulty of changing the system around it.
That is why it is useful. It is evidence that another organising logic can exist inside the public realm now.
This series has described what happens when services become factories: people are divided into processable needs and each organisation optimises its own part. August asked us to stop improving the candle and imagine the infrastructure required to produce the outcome.
Gateshead shows where that infrastructure can begin: with an organisation capable of staying with the reality of someone’s life; a relationship that holds context while specialists come and go; a worker trusted to respond to variation; a person with the power to say what matters to them; and leaders prepared to change the conditions when the system gets in the way.
For Mary, the outcome was not becoming better at navigating public services. It was no longer having to.