A Weavers Framework Assessment of the 10 Year Health Plan for England
“Fit for the Future” — Published July 2025 — Reviewed against the 76 Principles and Policies
April 2026 | David Sutton CITP MBCS | Southport Innovation Centre
This review applies the Weavers framework — its 76 principles and policies drawn from Appendix 3 (30 insights), Appendix 4 (8 key messages), Parts 1–3 of the image description, and the creation process — as a systematic analytical lens against the NHS 10 Year Health Plan for England published in July 2025. The review is structured in six sections: the plan at a glance; what it gets right; where the gaps are; the Weavers prescription; a summary assessment; and an overall verdict. Principle and policy numbers in parentheses refer to the 76 entries in the companion document.
SECTION 1
The Plan at a Glance
“Fit for the Future” is the UK Government’s 10 Year Health Plan for England, published 3 July 2025 following Lord Darzi’s 2024 investigation which found the NHS in a “critical condition”. At 168 pages, the plan sets out three strategic shifts and five technology bets as the framework for NHS transformation over the next decade.
Three shifts
Hospital to community; analogue to digital; treatment to prevention. P-24
Five big bets
AI, genomics, wearables, robotics, and data as the five transformative technologies. P-37
Scale
£29 billion real-terms increase in day-to-day spending over three years, backed by the Spring 2025 Spending Review.
Co-design
Over 250,000 contributions from the public, health staff and organisations via the Change NHS engagement process. P-27
Restructure
NHS England to be abolished, functions transferred to DHSC. ICBs merged into larger clusters with narrowed commissioning roles.
SECTION 2
What the Plan Gets Right
The Weavers framework finds genuine strengths here. A framework that only finds fault is not a useful analytical instrument.
Urgency is correctly named
“Major surgery, not sticking plasters” frames inaction as a choice with compounding consequences — not a neutral waiting position. The butterfly is correctly placed at the threshold. The plan understands it cannot simply manage the present into a better future. P-1, I-3
Three genuine structural shifts
Hospital to community, analogue to digital, and treatment to prevention are real structural moves, not repositioning. Each addresses a different layer of the broken clock: the care model, the technology architecture, and the design philosophy all require updating simultaneously. P-24
Citizens co-designed it
250,000+ contributions via Change NHS represents vertical cooperation from citizen level to strategy — the closest the NHS has come to the three-phase cooperation model, with Phase 3 (citizen-grounded validation) present at the design stage rather than bolted on afterwards. P-27, I-5
Five big bets broaden the AI stack
AI, genomics, wearables, robotics, and data. Partial but meaningful recognition that LLMs and foundation models alone are insufficient — the broader technology stack matters and has been named as a strategic commitment. P-37, KM-4
Federated Data Platform
NHS trusts collaborate while maintaining sovereignty over their own records. Each trust operates under its own governance processes with a national licence model. This is network architecture over tower architecture — the right structural instinct. P-11, P-12
Neighbourhood health teams
GPs, nurses, mental health, social care, pharmacists under one roof. Cooperation across professional silos designed into the physical and service architecture rather than left to aspiration. P-8, I-6
AI scribes as a frontline tool
Ambient AI transcription returning time to clinicians is the multiplier principle applied correctly: amplifying the expertise of the human practitioner, not replacing it. The fire given to someone who knows what they are holding. P-72, I-20, I-23
Ethical knowledge base in procurement
The plan signals AI governance frameworks with accountability, ethical review boards, audit trails, and incident reporting. Early recognition that attaching an ethical frame to AI deployment changes its quality. P-73
Co-design and test-and-learn signals
The commitment to co-designing technology with patients and staff, and the Scan-Pilot-Scale methodology, shows partial understanding of adaptive governance — design from the ground up before deploying at scale. P-25, P-26, I-15
SECTION 3
The plan’s genuine strengths cluster around two of the Weavers principles: the correct diagnosis of urgency and the recognition that the three shifts are structural, not incremental. The citizen co-design process is the most significant single addition relative to previous NHS strategies, because it is the closest the plan comes to the full three-phase cooperation architecture. These are the foundations on which a stronger strategy could be built.
Where the Gaps Are
Each gap below is mapped to the principle or policy it violates. The gaps are not random omissions — they share a common root.
Sovereignty unaddressed
The five big bets all depend substantially on US hyperscalers — Microsoft, Google, Amazon, Oracle — for cloud infrastructure and AI platform services. There is no 70/30 sovereignty threshold, no dependency cascade monitoring, and no named mechanism to ensure that as NHS AI capability grows, internal governance capability grows with it. Ethical sovereignty requires technical sovereignty: you cannot require modifications to systems you cannot inspect, and cannot protect patients from AI you cannot govern. The dependency cascade is not a theoretical risk: it is already in progress. The Stargate UK pause (April 2026) is one recent illustration — a project described at launch as strengthening sovereign compute for the NHS was paused by a foreign company’s commercial calculation about energy costs. That is what sovereign dependency looks like in practice. The NHS plan has no mechanism to prevent the same dynamic at the level of its AI models, its data platforms, or its clinical tools. P-1, P-2, P-3, P-4, P-5, I-10, I-11
No golden thread
168 pages of policy proposals exist without a traceable hierarchy connecting them. The test the framework carries — can you trace any major programme decision back to a specific policy, and from that policy to the mission? — fails at multiple levels in this plan. Policies are used as a compliance catalogue rather than as the connective medium between mission, vision, and programme portfolio. The NHS Children’s A&E analysis showed exactly this pattern: governance formally correct, strategically disconnected. P-21, P-23, P-25
Cooperation aspirational not structural
Neighbourhood teams and data sharing are described as ambitions. The plan calls for collaboration without mandating it. Insight 6 is precise: hoping that organisations will share is not a strategy. The initial resistance to mandated sharing is always identical across sectors, and the results always exceed expectation once it is required. Phase 1 cooperation — mandated within the sector — is absent. Phase 2 (across sectors) and Phase 3 (to citizen level) cannot follow without it. P-6, P-7, I-6
Multiplier test not applied
The plan deploys AI before examining what it will multiply. What is the quality of the data in individual NHS trusts? What are the information governance gaps? What processes are broken that AI will now amplify at scale? The question the framework requires — what is the quality of what we are multiplying? — is not asked. A strong AI platform over unexamined soil produces broken governance at network speed. P-14, P-15, P-43, P-47
Reorganisation without failure learning
NHS England is being abolished and ICBs restructured. The King’s Fund notes this is part of a consistent cycle of “redisorganisation” and asks why this plan will succeed when similar policies under previous governments have not. The framework requires: investigate the failure modes of the last reorganisation before designing the next one. The vine that grew in the ICB model has not been examined. RCA first, redesign second. P-28, P-30, P-59
No 2035 success statement
The plan’s own ambition is that by 2035 it will only have “begun narrowing health inequalities” — the King’s Fund called this “an unacceptably low ambition.” More fundamentally, no retrospective vision has been written first: no description of the NHS as it must be in 2035, from which strategy can be derived backwards. The clock still reads 2025. P-20, P-21, I-26
Determinism boundary absent
LLMs are being deployed in NHS clinical contexts — AI scribes, diagnostic support, coding tools — without a named governance framework for the specific property that makes them different from all previous AI: non-determinism. The same prompt produces different outputs. Clinical and evidentiary use requires governance built around this characteristic, not around the expectation that it will change. P-16, P-19, I-29
SDG alignment absent
The plan frames success entirely in terms of NHS productivity, adoption rates, and waiting times. The 17 UN Sustainable Development Goals — the internationally agreed definition of what a successful society looks like — appear nowhere. Every AI and transformation investment should be traceable to specific SDG targets. The question is not how much AI the NHS is deploying but whether the people it serves are better off on measures that matter. P-32, P-33, KM-2
AI stack narrower than stated
Despite naming five big bets, the technology governance in the plan focuses almost entirely on LLMs and foundation models. Edge AI, Local AI, and Tiny Micro AI — which matter for sovereignty and community-level access — receive no dedicated treatment. Agentic AI raises accountability governance questions not addressed by any current framework in the plan. The incomplete AI stack is a strategic gap. P-37, KM-4
Information quality not examined
The plan treats data as an asset to be unlocked — the National Data Library, Single Patient Record, Federated Data Platform. But the soil beneath the network is not examined. Data provenance, the processes that created it, the cultural practices that determined what was recorded and how — these are absent. Platform and knowledge substrate are distinct things that must both be present. P-43, P-44, P-45, P-47
SECTION 4
The common root of the gaps: the plan was built forward from the present rather than backwards from a designed future, and it was built by strategy people rather than systems people. The Weavers framework is a systems instrument. It finds what strategy documents miss precisely because it asks about the soil, the vine, the broken clock, and the network architecture — not only about the vision and the technology.
The Weavers Prescription: Seven Specific Additions
These seven additions would transform the plan from good strategy into great architecture. Each is grounded in the 76 principles and policies and traceable to the Weavers framework.
1
Write the 2035 address first Describe the NHS as it must be in 2035 — who is flourishing, which problems have been solved, what has changed in how care reaches the person without a car in Burscough, the community with the greatest health inequality, the frontline worker whose knowledge currently does not reach leadership. Then derive, backwards, what must be true in 2030, in 2027, and now. A strategy built by projecting forward from today is aimed at a world that will have moved on by the time it completes. The plan’s own admission — that by 2035 it will only have begun narrowing health inequalities — is the broken clock reading 2025. P-20, P-21, I-21, I-26
2
Apply the 70/30 sovereignty threshold across all AI procurement No more than 30% of strategic AI and technology capability may be externally sourced from any single foreign provider. Every AI procurement decision must pass the five-year test: after five years of this contract, will the NHS be more or less able to govern these systems internally? When sovereign capability is built on a foreign company’s investment decision, that company’s commercial calculus determines what NHS capability can and cannot do. The 70/30 threshold is not protectionism. It is the minimum condition for genuine choice. P-2, P-3, P-4, P-5, I-10, I-11
3
Map the soil trust-by-trust before deploying AI Assess data provenance, information quality, and the cultural and process conditions that created the data in every trust before AI is deployed on it. The surface symptom — a data quality metric, an anomaly in a report — is almost never the problem. It is the expression of something in the root system: a process that broke upstream, a definition that drifted, a cultural practice in which recording became compliance rather than genuine account. AI amplifies what it finds. The treatment is in the soil, not on the leaf. P-43, P-44, P-45, P-46, P-47
4
Make cooperation structural — mandate sharing across trusts Require sharing of AI models, operational data, failure modes, and innovations across all NHS trusts within the same sector. The initial resistance will be identical to every other sector in which mandated sharing has been introduced: ‘share with competitors? Impossible.’ The results will exceed expectation within 18 months. Then extend to Phase 2: cross-sector cooperation with utilities, local government, and social care. The NHS already knows that the best solution to a problem in an emergency department may exist in a water network’s anomaly detection system — but this knowledge cannot be accessed by accident. P-6, P-7, P-8, I-5, I-6, I-7
5
Govern the determinism boundary for all LLM clinical use Establish a named governance framework for the specific property that distinguishes LLMs from all previous AI systems: non-determinism. The same prompt produces different outputs. This is not a defect to be patched — it is a design characteristic that requires governance built around it. Any use of LLMs in clinical settings, evidentiary contexts, or as the basis for treatment decisions requires: a stated governance framework for non-deterministic outputs; human-in-the-loop requirements at consequential decision points; and audit trails that account for output variability. P-16, P-19, I-29
6
Investigate the last reorganisation before designing the next The vine that grew in the ICB model — the barriers that prevented its stated cooperative purpose from being achieved, the information flows that broke between its structures and the frontline, the reasons why integration remained aspirational — must be examined before the next structure is designed. Root cause analysis of the ICB model should precede the abolition timetable, not follow it. The Post Office Horizon pattern, the Birmingham City Council pattern, and the NHS restructuring patterns of the last 25 years share a single root: failures hidden because the vine was never examined. RCA first. Redesign second. P-28, P-30, P-57, P-59, I-14
7
Apply the blue flower test to every technology deployment Before any technology deployment goes live, ask: does it work for Oliver — seven years old, epilepsy, no family car, Burscough? Does the NHS App work for the patient without a smartphone, reliable broadband, or digital literacy? Does the neighbourhood health centre work for the person who cannot take time off work to attend in the hours it is open? Does the AI diagnostic tool perform equally across demographic groups, or does it perform less well for the populations who most need it to perform well? The red flower will be fine. Design for the blue flower. That is not an inclusion exercise. It is the design constraint. P-24, P-25, I-4, I-16, I-15 — Appendix 9 (Southport Children’s A&E case)
SECTION 5
Summary Assessment Table
The table below summarises framework findings across key dimensions. ✓ Strong △ Partial ✗ Gap identified
Dimension
Assessment
Framework ref
Urgency and pace
✓ Inaction correctly framed as compounding risk, not a neutral default
P-1, I-3
Three strategic shifts
✓ Hospital to community, analogue to digital, treatment to prevention are genuine structural moves
P-24
Citizen co-design
✓ 250,000+ contributions — most significant vertical co-design in any NHS strategy
P-27, I-5
Five big bets (AI stack)
△ Technology scope broadened beyond LLMs but governance detail has not followed
P-37, KM-4
Federated Data Platform
✓ Cooperative data network with local sovereignty — correct structural instinct
P-11, P-12
AI scribes and frontline AI
✓ Multiplier applied to clinical expertise — augmentation, not replacement
P-72, I-20
Sovereignty — 70/30 threshold
✗ Five big bets rely on US hyperscalers — no threshold, no monitoring mechanism
P-2, P-3, I-10, I-11
Sovereign dependency cascade
✗ Cascade already in progress — no mechanism to track or limit outsourcing of AI capability
I-10
Traceable policy hierarchy
✗ Policies function as a compliance catalogue, not a connective chain from mission to action
P-21, P-23
Cooperation — mandated sharing
✗ Collaboration described as an ambition — no structural mandate across trusts or sectors
P-6, P-7, I-6
Pre-deployment data quality
✗ AI deployed without assessing the quality of the data it will operate on trust-by-trust
P-14, P-15, P-43
Reorganisation failure analysis
✗ No independent analysis of ICB model failure modes before designing successor structures
P-28, P-30, P-59
2035 outcome-based vision
✗ No success statement written from 2035 — strategy optimises for the present it can see
P-20, I-21, I-26
LLM non-determinism governance
✗ LLMs deployed clinically without governance for non-deterministic output variability
P-16, P-19, I-29
Adaptive governance
△ Scan-Pilot-Scale is present but sits within a fundamentally linear delivery architecture
P-25, I-13
SDG outcome alignment
✗ Success measured by adoption rates and waiting times — SDGs absent as an outcome framework
P-32, KM-2
Information quality — soil
✗ Data treated as an asset to unlock — provenance, quality, and cultural origins not examined
P-43–47
Inclusion test at deployment
✗ No test applied for populations with the most complex access requirements
P-24, I-4, I-16
SECTION 6
Overall Verdict
The plan is strong infrastructure policy and weak transformation strategy. It will succeed at shifting care settings, deploying AI scribes, expanding the NHS App, and building digital plumbing. It will not, on its current architecture, produce the three-dimensional cooperative transformation the framework identifies as necessary, the adaptive governance that navigates non-linear transitions, or the human flourishing measurement that would tell whether any of it is working.
The plan reads the present clearly and projects forward optimistically. The Weavers discipline requires something different: stand in 2035 and read backwards. Ask what must be true about the NHS in 2035 — not in terms of technology adoption rates, but in terms of whether Oliver, seven years old, epilepsy, no family car, Burscough, is receiving care that meets the organisation’s stated mission. Then ask what must be built now, in what sequence, for that to be achievable.
The butterfly is correctly placed at the threshold. The urgency is real. The three shifts are genuine. The citizen co-design is the most significant structural addition relative to any previous NHS strategy. These are the right foundations. But the soil beneath the network has not been examined, the vine between the allocated governance and the actual governance has not been named, the clock on the tower still reads 2025, and the architecture of the digital infrastructure — built on foreign company investment decisions — is the tower model, not the network model.
The most important single addition that would shift the plan’s architecture is also the simplest: write the 2035 address first. Describe what the NHS will have become — not what it will have built. Then test every programme decision against it. The network grows stronger with every season. The tower resets. The choice of architecture is the defining decision of this era. The choice is ours.
References and Framework Notes
Source document: Fit for the Future: 10 Year Health Plan for England. UK Government / NHS England, 3 July 2025.
Framework: Weavers and The Web — Extended Version v24, David Sutton CITP MBCS, Southport Innovation Centre, March 2026.
P-N: Principle or Policy N from the 76-entry companion document (Weavers Principles and Policies).
I-N: Insight N from Appendix 3 of the Weavers document (30 insights from the UK Industry 4 Transformation Strategy).
KM-N: Key Message N from Appendix 4 (8 key messages applying the Weavers lens to the Industry 4 model).
App. 9: Appendix 9 of the Weavers document — the NHS Children’s A&E case study (Southport/Ormskirk), which first demonstrated the golden thread diagnostic and the blue flower principle in a live NHS context.