Dr Graham Turnbull, Partner at Reinvigoration, will deliver a keynote at the Housing Cymru Conference in Cardiff on 1 July 2026. This article sets out the argument he will make: that the demands of the decade ahead converge on a single operating model, that operational capacity is therefore the binding constraint, and that simplifying the operation is the most effective response, because it releases the capacity to deliver and creates the foundation on which artificial intelligence can be applied effectively.
The demands on social housing in Wales have rarely been greater, and they have rarely arrived so closely together. Over the decade ahead, registered social landlords are expected to decarbonise their stock under the Welsh Housing Quality Standard 2023 (WHQS 2023), respond to damp, mould and other hazards within defined timescales, sustain financial resilience under combined cost pressure, and adopt new technology and artificial intelligence (AI), all whilst continuing to deliver the repairs, lettings, complaints handling and tenancy services that tenants rightly expect.
It is tempting to treat each of these as a separate programme, owned by a different director, with its own plan, governance and reporting. In our experience, that is the first mistake.
None of these demands are delivered in isolation. Every one of them runs through the same operation and is carried by the same teams. The operative who attends a hazard within the new timescale is the same resource the organisation relies on to deliver its retrofit programme. The data that evidences compliance is produced by the same systems that schedule planned works and resolve complaints. When the demands are added together, what decides the decade is not the organisation's ambition or its funding, but whether it has the operational capacity to deliver all of it through one operation, at once.
That is the argument I will put to housing leaders at the Housing Cymru Conference in Cardiff on 1 July, and it is the case I want to set out here.
What decides the decade is whether the organisation has the operational capacity to deliver all of it through one operation, at once.
The demands of the decade converge on a single operating model
Consider the scale of what is being asked, from an operational standpoint rather than a policy one.
The Welsh Housing Quality Standard 2023 came into effect in April 2024 and must be met in full by 31 March 2034. Its energy efficiency and decarbonisation requirements, set out in Part 3, will not be included in compliance assessments until March 2027, but the milestones in between are demanding. According to the Welsh Government, landlords were required to assess their stock and submit a compliance policy by March 2025, must develop a Target Energy Pathway for each of their homes by March 2027, and are expected to reach an Energy Performance Certificate (EPC) C standard across their stock by 2030 on the way to full compliance in 2034.
The early evidence shows how much ground remains. According to the Welsh Government's first annual statistics, covering the position as of 31 March 2025, 48% of social homes met all relevant elements of the standard, whilst 47% did not yet meet it. These figures do not yet include the decarbonisation requirements, which are widely recognised as the most demanding part of the standard.
The Welsh Government's own evaluation of WHQS 2023, conducted by Alma Economics and published in 2026, is candid about why. It found that the energy efficiency and decarbonisation requirements were seen as the most challenging aspect of the standard, and that whilst the rationale for high standards was widely supported, the pace and scale of ambition were, in some cases, perceived to outstrip available capacity. The constraints it identified are revealing: financial pressure, uncertainty around decarbonisation funding, shortages of trained and accredited personnel, and the administrative complexity associated with compliance reporting.
Decarbonisation is only one of the demands converging on the operation. From 1 April 2026, following a consultation prompted by the death of Awaab Ishak in England, the Welsh Government has updated the WHQS to set out clear timescales for investigating and remedying hazards, including damp and mould. As the Welsh Government has set out, social landlords must now publish their response times and report on performance as part of the WHQS compliance return, alongside their existing legal duty under the Renting Homes (Wales) Act 2016 to keep homes fit for human habitation. This is a significant operational change. It requires not only a faster response on the ground, but the systems, data and clear ownership needed to evidence that response consistently and to report it credibly.
Set these alongside the financial pressure of investing in stock whilst holding down cost to serve, and a workforce that is already stretched and short of the skills the retrofit programme requires, and the pattern becomes clear. The conference programme itself recognises this, framing the decade ahead around the need for knowledge, a robust supply chain and a skilled workforce. Each of these demands is real, each is legitimate, and each land on the same operating model and the same people. That is what makes operational capacity the binding constraint of the decade.
Each of these demands lands on the same operating model and the same people. That is what makes operational capacity the binding constraint of the decade.
Where capacity is lost
If capacity is the constraint, the practical question is where it goes, and the answer is that a great deal of it is consumed by complexity that has accumulated over years rather than by the work itself.
Repairs is the clearest illustration, though it is one example among many. A job that should take a few days can take far longer when it crosses several disconnected systems and passes through multiple hands between triage, scheduling and completion, with each handoff adding delay, duplicated effort and the risk of information being lost along the way. The cost of this is significant but rarely visible, because it hides in additional handling time, rework, reconciliation between systems and the effort required to produce consistent management information. It sits beneath the headline figures without ever appearing as a named line item.
The sector's own data shows the scale of the prize. According to figures reported across the sector, social housing providers in the UK spent a record £8.8 billion on repairs and maintenance in 2024, 13% higher than the previous year and 55% higher than five years earlier. Research by the Ministry of Housing, Communities and Local Government's Local Digital programme estimates that around £400 million is wasted every year on maintenance that is not delivered properly, for just two services, repairs and allocations, much of it traceable to poor and fragmented data. Social landlords have, as Housing Digital observes, historically relied on separate systems for repairs, rent, customer contact and asset data, which makes it difficult to form a single, accurate view of a tenant or a property.
Repairs is illustrative rather than unique. The same capacity is consumed, and the same capacity can be released, across lettings and voids, complaints handling, income collection, and the scheduling of planned and compliance work. In our work with housing associations, simplifying these end-to-end journeys has typically reduced lead times and cost to serve by around a fifth, without additional headcount, by removing unnecessary variation, clarifying ownership and reducing the number of handoffs and systems a piece of work has to cross. The capacity released in this way is precisely the capacity the decade demands.
Social housing providers in the UK spent a record £8.8 billion on repairs and maintenance in 2024, with around £400 million wasted each year on maintenance that is not delivered properly.
Why technology alone will not release that capacity
Faced with this pressure, it is natural to look to technology, and to AI in particular, to create the headroom. The expectation is understandable, and the conference rightly gives technology a prominent place. My experience suggests that something more careful is needed, because the value of technology depends on the state of the operation beneath it.
The sector's own adoption pattern bears this out. According to the National Housing Federation's 2025 survey of the sector, well over nine in ten housing organisations had experimented with some form of AI, yet fewer than one in five had embedded its use or put formal governance and training in place. The typical uses were modest and practical, such as drafting correspondence, summarising documents, and supporting early triage in repairs, complaints and arrears, and they often emerged informally as teams looked for ways to cope with capacity pressure. The more advanced applications, particularly predictive and prioritisation tools, remained exploratory rather than embedded, constrained, as the evidence consistently shows, by uneven data quality, skills gaps and uncertainty about long-term ownership and assurance.
This is the heart of the matter. Where a service operates in several different ways depending on team, system or the legacy structure that still shapes how work flows, automation does not remove that variation. It replicates it, leaving the organisation with multiple automated pathways, each of which has to be maintained, governed and explained. The expected efficiency gain is diluted, and the burden simply shifts into a different form. Applied to a fragmented operation that is accountable for vulnerable residents, technology can also introduce real risk, because a decision that cannot be traced cannot easily be assured.
This is the familiar pattern of a confident digital front door sitting on top of a fragmented operation behind it, and it is well established across the wider public sector. Reinvigoration's analysis has drawn on research by Socitm finding that whilst 72% of councils had invested in digital platforms, fewer than 25% had integrated them with their back-office operations. Technology has not failed these organisations. It has exposed the limit of the operation it was asked to run on.
Automation applied to a fragmented operation does not remove the variation it replicates it.
The simple unlock: one act of simplification, two returns
The implication is not to do less, and it is certainly not to abandon technology. The most effective response is to simplify the operation first, because simplifying the end-to-end operation produces two returns at the same time.
The first return is capacity. Removing unnecessary variation, clarifying who owns each outcome, and reducing the handoffs and systems that work has to cross releases time and effort that existing teams can redirect to the priorities that matter, whether that is the retrofit programme, the new hazard timescales, or the day-to-day repairs and tenancy services on which tenants depend. This is capacity created from within the organisation, rather than capacity bought through additional headcount that the sector can neither easily afford nor readily recruit.
The second return is a foundation for technology. The same act of simplification produces the clean, consistent processes, the clear ownership of decisions and exceptions, and the trustworthy, well-defined data that AI and automation need in order to work reliably and at scale. As the sector's own commentators have observed, value from AI does not come from adoption alone, but from discipline, evidence and trust. Simplification is how an organisation builds that discipline into the operation, so that when technology is applied it accelerates good work rather than encoding existing complexity.
This is the logic of our Simplify4Scale® methodology, which was designed for exactly this challenge. It begins by making operational reality visible, mapping how work actually flows through the organisation rather than how it is documented, then simplifies before scaling, and builds the internal capability so that improvement is sustained long after external support has ended. The sequence matters, because clean foundations laid before the next wave of change are what allow that change, whether driven by regulation, technology or future restructuring, to deliver as intended.
Simplifying the operation releases the capacity to deliver the decade and creates the foundation on which AI can be applied well. Both returns come from the same act.
What this means for housing leaders in Wales
For chief executives, chief operating officers and directors of operations across the Welsh housing sector, the conclusion is practical rather than abstract. The decade ahead will be demanding regardless of the choices made, but the organisations that fare best will be those that recognise capacity, rather than ambition or even funding alone, as the binding constraint, and that act on it by simplifying the operation before layering further change on top of it.
The path forward starts with operational evidence rather than strategic aspiration: an honest, end-to-end view of how work flows today, where the delays, duplication and divergent definitions sit, and where capacity is being consumed without adding value. From that evidence, simplification can begin, and from simplification comes both the capacity to deliver the decade and the foundation to apply technology well. The two are not separate projects competing for the same stretched teams; they are two returns from the same act of simplification.
I look forward to making this case in person at the Housing Cymru Conference on 1 July, and to the conversations that follow. If you would like to discuss what it would mean for your organisation, our white paper, The New Shape of Public Sector Transformation, examines where operational complexity comes from, why it now prevents change from delivering, and what a proven approach to addressing it looks like in practice.
