Thought leadership, case studies, industry news and forward-looking analysis — written by engineers and practitioners who work in the same sectors you do.
Most infrastructure organisations can tell you how many assets they own and when they were last inspected. Far fewer can answer the question that actually matters: are those assets healthy enough to keep doing their job — today, and in the years ahead? This article explains the seven data groups, the Asset Health Index methodology, and a practical three-phase approach to building a structured asset health picture — applicable across all regulated sectors globally.
The structural knowledge problem facing infrastructure teams, and a practical path to fixing it through AI-assisted lessons capture. First published in the Institute of Water Magazine, Winter 2025.
The seven data groups, the Asset Health Index methodology, and a practical three-phase approach — applicable across all regulated infrastructure sectors globally.
Six clause changes that matter in practice — decision-making framework, data governance, organisational knowledge, life cycle management — plus the transition timeline and where to start.
Half the knowledge in UK water infrastructure will retire in the next few years. Recruitment solves next decade’s problem — here’s what can bridge the gap now.
Why replacing the worst-condition assets first is usually the wrong call — and the three-layer framework mature organisations use instead.
Subscribe below to be notified when new articles are published.
How an operator moved from manual spreadsheet-based condition recording to a structured, ISO 55001-aligned evidence pack without changing their inspection process.
How ProjectHindsight turned closeout reports from real UK infrastructure programmes into a live, searchable knowledge base.
We are documenting live pilot results as they emerge.
UKWIR published its “Common Definition & Calculation of Asset Health” framework — the sector’s first practical standard for consistent asset health assessment across UK and Ireland.
The white paper commits to new mandatory resilience standards and a new single regulator with an embedded Chief Engineer overseeing asset health. What infrastructure operators need to know.
Ofwat’s December 2024 Final Determinations include a specific roadmap for enhancing asset health understanding across the 2025-27 period, building toward PR29.
The UK has around 300,000km of buried water mains. Most have never been physically inspected. How AI-based condition inference could create a probabilistic health picture of the buried network.
Around half of UK water engineers are over 50. What AI can realistically do to capture the knowledge that cannot be found in any document — and what it genuinely cannot replace.
We write about where infrastructure engineering is heading, not just where it is now.
Everything published here connects to real engineering challenges — not generic tech commentary.
Lessons learned, knowledge capture frameworks and how AI is changing what is possible.
ISO 55001, condition grading, deterioration modelling and remaining useful life.
Whole life cost modelling, intervention prioritisation and regulator-ready investment cases.
Ofwat, PR24/PR29, the Water Reform Bill, ISO 55001 licence conditions and what changes mean in practice.
Responsible AI, human-in-the-loop design, EU AI Act obligations and deploying AI where safety matters.
Asset health submissions, UKWIR research and the practical realities of UK water infrastructure delivery.
Rail, highways, NHS estates, energy networks and aviation — where the same asset management challenges apply.
Digital twins, buried pipeline condition inference and what infrastructure engineering will look like in 2035.
I have spent 17 years delivering infrastructure projects. Nigeria’s oil and gas sector first, then UK water and energy. Different countries, different regulators, different asset types — but one thing that stayed the same everywhere I worked: the knowledge that could have prevented the last failure was already written down somewhere. It just was not where anyone could find it when the next project started.
Post-project lessons get written. They get filed in a SharePoint folder with a sensible name. And then they get completely ignored on the next scheme — not because the next team is careless, but because nobody told them the folder existed, or the search does not work, or the document is 80 pages long and the relevant bit is buried in appendix F.
The result is predictable. The same consenting delays happen on the same type of scheme. The same supply chain problem surfaces at the same programme stage. The same commissioning fault occurs on the same class of plant. We call them lessons learned, but a lesson only counts if someone learns it — and for that to happen, it has to reach the people who need it, at the moment they need it, in a form they can act on.
It would be easy to frame this as an attitude problem — teams not bothering to document what went wrong, or not taking the time to read what previous teams wrote. But that is not what I observed. Most of the engineers I worked with were conscientious. They wrote detailed closeout reports. They ran post-project reviews. They produced lessons-learned registers.
The problem is structural. The way knowledge is currently stored — in documents, organised by project, sitting in shared drives — makes it almost impossible to retrieve at the point it would be useful. When you are starting a new pumping station project, you do not search for “pumping station closeout report 2019.” You search for “what went wrong at commissioning on schemes like this one.” Those are different questions, and traditional document management systems cannot answer the second one.
UK infrastructure has another dimension to this problem that makes it more urgent: the workforce is ageing fast. Research consistently places around half of UK water industry engineers over 50. The same pattern applies across rail, highways, energy, and NHS estates.
This matters because a significant portion of the knowledge that keeps infrastructure running safely is not written down anywhere. It lives in the heads of experienced engineers — the ones who know from memory which soil type causes which problem, which contractors have which weaknesses, which type of asset tends to fail first in which conditions. When those engineers retire, that knowledge retires with them.
Ofwat’s December 2024 PR24 Final Determinations and the January 2026 Government White Paper “A New Vision for Water” both identify asset knowledge and condition management as critical national risks. The sector is being asked to deliver the most capital-intensive regulatory cycle in its history against a backdrop of an ageing workforce and a documented knowledge loss problem.
What AI can do well: read large volumes of unstructured text and identify patterns, extract structured information from documents written to be read by humans rather than searched by machines, and surface relevant past information when someone asks a question about a new situation.
What AI cannot do: make engineering judgements, replace professional experience, or be trusted to produce final outputs without human review. This is the correct design for any AI system operating in a regulated, safety-critical context. AI should augment engineering judgement, not bypass it.
The approach I took when building ProjectHindsight was to make human review mandatory at every step. AI extracts a lesson from a document — but an engineer reviews and approves it before it enters the knowledge base. AI ranks the relevance of past lessons — but the project team decides which ones to act on.
At the end of a project, an engineer uploads the closeout report. The AI reads it and extracts structured lessons — each tagged with a title, root cause, impact, recommendation, and category. The engineer reviews and approves them into the knowledge base.
When the next project starts, the project manager searches by topic — not by document name. “Consenting on chalk geology schemes.” “Commissioning issues with horizontal centrifugal pumps.” The AI retrieves the most relevant past lessons, ranked by match score, regardless of which project they came from.
The knowledge to prevent most infrastructure project failures already exists somewhere in the organisations that suffered them. The problem has never been the absence of knowledge — it has always been access. That problem, finally, is one that AI can genuinely help to solve.
Most infrastructure organisations can tell you how many assets they own, where they are, and when they were last inspected. Far fewer can answer the question that actually matters: are those assets healthy enough to keep doing their job — today, and in the years ahead?
That distinction is the difference between an asset register and an asset health system. One tells you what you have. The other tells you what condition it is in, how much useful life remains, how reliable it has been, whether it meets its compliance obligations, and — critically — how much risk it represents to the people and services it supports.
UK Water Industry Research published its “Common Definition and Calculation of Asset Health” framework in April 2026 — the sector’s first practical standard for consistent asset health assessment. It defines asset health as: “an asset’s ability to perform its intended function now and in the future.” The same definition applies equally to a water treatment works, a railway bridge, a hospital boiler room, or an airport runway.
A proper asset health system is built from multiple sources of evidence, structured into a consistent framework across seven data groups. Based on the UKWIR framework and equivalent methodologies used globally across energy, rail, highways and aviation sectors:
| # | Data Group | What it captures | Why it matters |
|---|---|---|---|
| G1 | Asset Identity | Unique ID, name, class, location, GPS | The foundation everything else sits on. Wrong hierarchy = misleading analysis. |
| G2 | Technical Condition | Condition grade, material, defect records | Most direct evidence of health — but only if structured, not buried in a PDF. |
| G3 | Age & Life Cycle | Installation year, design life, consumed life % | An asset at 120% of design life is very different from one at 60% — even with the same grade. |
| G4 | Performance & Reliability | Failure history, operational status | An asset that has failed three times is not healthy, even if the last inspection was Grade 2. |
| G5 | Compliance | Statutory inspection dates, regulatory obligations | A non-compliant asset is an immediate liability regardless of its condition grade. |
| G6 | Asset Criticality | Consequence of failure, redundancy, population served | Turns a health score into a risk score. Same grade, very different priority. |
| G7 | Data Governance | Completeness score, confidence level, last verified date | A score is only as trustworthy as the data behind it. |
In almost every infrastructure organisation attempting structured asset health assessment, the first finding is the same: the data exists, but not in a usable form. Condition grades sit in PDF survey reports. Installation dates sit in procurement records. Failure history sits in work orders. The problem is rarely an absence of data — it is fragmentation. The first job is assembly and standardisation, not new surveys.
The data groups above combine into a single Asset Health Index (AHI) — a composite score from 0 to 100. The methodology below is derived from the UKWIR April 2026 framework and is applicable across all regulated infrastructure sectors globally.
The AHI tells you how healthy an asset is. It does not tell you how important it is that the asset stays healthy. A Grade 3 pump with no standby is a very different priority to a Grade 3 pump with full redundancy. The AHI is multiplied by a Criticality Multiplier to produce a Risk Score on a 0–300 scale.
An asset with an AHI of 40 and Tier A criticality produces a Risk Score of 120. The same AHI on a Tier C asset produces 40. Both need attention — but the priority is unambiguously different, and this is how asset health data connects directly to capital investment decisions.
The most common mistake is waiting until all seven data groups are complete before producing any output. A partial picture with clearly flagged gaps is more useful and more honest than waiting years for perfect data. Work in three phases:
Every conversation about the UK water sector's ageing workforce eventually lands on the same framing: we are about to lose a lot of experienced people, and we need to recruit and train our way out of it. That framing is not wrong. But it treats the crisis as a headcount problem, when for asset management specifically, it is something narrower and more urgent — a data problem, and one with a closing window to fix.
Research commissioned by the Engineering Construction Industry Training Board puts the UK water treatment contractor workforce at close to 50% aged over 50, with only around 8% under 24. CIWEM's 2026 "Your Future" report — drawing on its Member Career Survey and the latest Energy & Utilities Skills figures — estimates the sector needs an additional 43,700 workers by 2030, a 36% increase, just to replace retirees and deliver AMP8's £104 billion investment programme.
Sit with that second figure for a moment. A 36% workforce increase is not a training pipeline problem you solve by running more apprenticeships, though the sector is rightly doing that too — Water UK and government have committed to 50,000 new recruits by 2030, including 5,000 apprenticeships. It is a statement that a very large share of the people who currently hold institutional knowledge about specific assets will be gone within a few years, whoever replaces them.
The Murray McIntosh Water Industry Labour Report — one of the sector's most widely cited workforce studies, now informing the Independent Water Commission's own review — has repeatedly surfaced a specific and uncomfortable pattern: a substantial share of engineers report considering a move to another sector well before formal retirement age. That, combined with the ECITB and CIWEM figures above, points to an honest conclusion: knowledge loss is not a future event scheduled for someone's 65th birthday. It is already happening, gradually, every time an experienced engineer moves on.
What actually leaves is rarely written down anywhere formal, and it is not only about condition. It is knowing that a particular reservoir's true condition doesn't match its last survey grade, because of something observed on site three years ago that never made it into a report. But just as often, it is knowing that an asset with an unremarkable condition grade actually carries far more consequence than its grade suggests — because of a downstream dependency, a lack of redundancy, or a customer impact that only someone who has worked that network for years would know to flag. That second kind of judgement, about criticality rather than condition, is if anything harder to replace, because it rarely gets written down at all, even informally. None of this is negligence — it's simply how experienced engineering judgement has always worked, held in memory rather than in a system, because for most of the sector's history there was no practical alternative.
A skills shortage affects every function in a water company. But it hits asset management in a way that is structurally different from, say, customer service or finance. A defensible investment decision depends on two kinds of judgement, not one — how healthy an asset actually is, and how much it would matter if it failed. Both are captured, today, as a number on a form, with the reasoning behind that number left in the assessor's head rather than the record itself.
This connects directly to the seven data groups any structured asset health system needs, discussed in an earlier piece on this page. Group 6 — asset criticality — is exactly the judgement most at risk here, since it depends on knowing a network's real dependencies and consequences, not just reading a condition report. Group 7 — data governance — is the practical response: it requires a named owner and a confidence level for every recorded score, precisely because of this workforce problem. A criticality score or condition grade recorded by someone with fifteen years on that specific network carries a different confidence level than the same score recorded by someone six weeks into the role, even if the number on the form is identical.
Every recruitment and training initiative described above is genuinely necessary, and none of it is fast enough on its own. A graduate hired this year under Water UK's expansion plans will not carry thirty years of institutional knowledge about a specific asset base for a long time yet, however well they're trained. The sector is, in effect, racing to build a new generation of judgement before the current generation's judgement retires with them — and losing that race in places, based on how consistently "skills and recruitment" now tops member surveys as the industry's leading concern.
The uncomfortable trade-off is this: recruitment and training solve the problem over a ten-to-fifteen-year horizon. The knowledge loss is happening over a two-to-five-year horizon. Something has to bridge that gap, and it cannot be more people, because there aren't more experienced people to hire — the whole sector is drawing from the same shrinking pool.
Not replacing the retiring engineer — that isn't possible on the relevant timeline. The realistic goal is capturing enough of what they know, in a structured, searchable form, before they leave, so the judgement behind a condition grade survives even when the person who made it doesn't stay. That is a data and knowledge-management problem, not a recruitment problem, and it can be started this year rather than waiting for a ten-year pipeline to mature.
The sector's recruitment and training response to this crisis is real, well-funded, and necessary. But it is solving next decade's problem. The knowledge walking out the door this year needs a different kind of answer — one built around capturing and structuring what experienced engineers know, before the option to do so closes for good.
Ask five water companies to define "asset health" and you will likely get five workable, defensible, and subtly different answers. That has been the practical reality across the sector for years — not because companies are careless, but because no shared standard existed to align around. UKWIR's new framework is the sector's first serious attempt to close that gap.
UKWIR's own framing is direct: management of asset health has long been recognised as a concern precisely because there was no agreed approach to quantifying and valuing it. Several definitions existed with a degree of overlap, but the lack of alignment made cross-company comparison, sector-wide benchmarking, and consistent regulatory reporting genuinely difficult.
The framework recommends a single definition for sector-wide adoption: asset health is an asset's ability to perform its intended function now and in the future. It is a deliberately simple definition, designed to be measured consistently rather than debated endlessly.
Three practical building blocks: a common asset class hierarchy and taxonomy, agreed good practice in how asset condition is assessed and measured, and good practice in how monitoring and intervention decisions get made from that data. Importantly, UKWIR designed it to build on what companies already have in place, rather than requiring a rip-and-replace of existing systems — companies are encouraged to map their existing asset health structures onto the common framework, making genuine cross-sector comparison and shared learning possible for the first time.
The framework was not developed in a vacuum, and its implementation is already surfacing real friction. Ofwat's own parallel work on a common asset hierarchy — developed with Turner & Townsend — considered UKWIR's ten-tiered hierarchy and, after discussion with Ofwat, settled on a six-tiered structure instead for their specific regulatory purpose. That is not a contradiction so much as an honest signal: a framework built for internal asset management decisions and one built for regulatory reporting do not automatically serve both purposes equally well. UKWIR's own research reached the same conclusion directly, recommending the standard definition align more closely with the regulatory definition precisely because the same definition was not effectively serving both uses.
The framework arrives at a genuinely useful moment. It directly supports the Government's January 2026 White Paper, providing the kind of consistent metrics the White Paper's proposed new regulatory requirements will need to actually function. For any organisation building or refining an asset health system today — regardless of sector — the practical lesson from UKWIR's work is the same one this platform is built around: get your definition and your data groups aligned first, because the framework only earns its value once companies actually map their real data onto it.
Most engineers working in UK water will already know the headline: Ofwat is being abolished. A new single integrated regulator will absorb the water-related functions of Ofwat, the Drinking Water Inspectorate, the Environment Agency, and Natural England into one body. But the detail that matters most for asset management specifically is what that new regulator is actually built to do differently.
The new regulator embeds a Chief Engineer at its core, leading hands-on infrastructure inspection and a dedicated engineering team with real technical capability, not just economic oversight. The White Paper explicitly introduces an "MOT" approach to water infrastructure — maintenance checks on pipes, pumps, and treatment works, moving the sector away from operator self-monitoring toward a more proactive, company-specific supervisory model.
The White Paper is direct about the scale of the challenge: around 60% of water mains were built before 1981, and approximately 13% are over 100 years old. That statistic alone explains why "improved mapping of asset health" and statutory resilience standards sit near the top of the reform agenda.
For anyone building capital investment cases, the practical reforms matter as much as the institutional ones. The White Paper commits to clearly defining and ring-fencing capital maintenance expenditure so it can only be spent on maintaining assets, with separate allowances for capital maintenance, operating expenditure, and enhancement capital expenditure at every future price review, starting with PR29. The current system of price control redetermination through the Competition and Markets Authority is also being replaced with a more focused appeals process, aligning water more closely with how the energy sector already operates.
Underperforming companies face a new Performance Improvement Regime, giving the new regulator the power to intervene earlier and more proactively rather than waiting for a scheduled price review to address failing asset health. Combined with the end of self-monitoring, this is a genuine shift from periodic, retrospective assurance toward continuous, evidence-based oversight — which only works if the underlying condition and criticality data is actually good enough to support it.
This is not an overnight change. The White Paper itself is the opening move: a Transition Plan follows in 2026, a Water Reform Bill is expected in late 2026 or early 2027, and the new integrated regulator is targeted for establishment in mid-to-late 2027. For organisations doing investment planning today, the practical message is to start building the asset health evidence base now — structured, defensible, and ISO 55001-aligned — rather than waiting for the new regulator to ask for it.
PR24 delivered the largest single funding settlement in the water sector's history — a total spending allowance of £104 billion, split between £60 billion of base expenditure and £44 billion of enhancement expenditure, an 18% increase on the draft determinations published five months earlier. Buried within that settlement is a quieter but consequential commitment: a structured plan to fix a genuine gap in how well the sector actually understands its own asset health.
Ofwat collected robust asset condition data from companies on water mains, sewers, and bioresources assets through the PR24 process itself. But that same process surfaced an uncomfortable finding: condition data is not collected in the same way, or to the same extent, across the rest of the asset base — treatment works, service reservoirs, pumping stations, and more. Some asset classes have decent, comparable data. Many do not.
Ofwat uses historical workload and condition data to determine what its existing base expenditure allowances should already be covering — the regulator's phrase for this is "what base buys." Without comparable condition data across the full asset base, it is genuinely difficult to tell whether a sector-wide asset health issue is real and needs additional funding, or whether it's already meant to be covered by existing allowances. Better data isn't a bureaucratic nicety here; it is the difference between the regulator having grounds to fund a genuine gap and companies being told to find it within what they've already got.
The roadmap sets out a defined two-year working period. Ofwat is engaging directly with water companies and wider stakeholders through workshops and working groups to build a common asset inventory, identify a prioritised subset of assets to focus data collection on, and agree consistent methods for collecting condition and workload data (repair, refurbish, replace, or other intervention types) across that subset. The explicit goal is deciding, with real evidence, whether additional base expenditure allowances are needed to address sector-wide asset health issues before the next price review, PR29.
Worth noting directly: Ofwat's own update papers on this workstream explicitly reference ISO 55001:2024 as a framing standard. This is not a coincidence — a regulator asking "do you have comparable, structured, confidence-rated condition data across your asset base" is asking almost exactly the question ISO 55001's data governance and risk clauses are built around. Companies already building toward genuine ISO 55001 alignment are, in practice, already building toward what this roadmap will eventually ask of them.
The practical lesson for any asset manager reading this: the assets where your condition data is weakest today are precisely the ones a regulator-led review is most likely to focus on next. Closing that gap proactively, ahead of being asked, is the stronger position to be in before PR29 conversations begin in earnest.
Somewhere under almost every UK street runs a pipe that has never once been physically inspected, and in most cases, never will be. This isn't negligence — it's basic economics. Physically inspecting hundreds of thousands of kilometres of buried pipe, one length at a time, was never going to happen at the pace or cost anyone could justify. That reality is precisely why the case for a different approach — probabilistic, AI-based condition inference rather than universal physical inspection — is worth taking seriously.
The UK's water mains network runs to roughly 300,000km of buried pipe. The National Audit Office has calculated that at the sector's current replacement rate, renewing the entire network would take around 700 years — against a sustainable replacement cycle closer to 125-160 years. England and Wales lose an estimated 3 billion litres of treated water a day to leaks: enough to fill 1,200 Olympic swimming pools, every single day.
Around a third of the UK's water mains network is now plastic, added over decades of repair and expansion alongside the older metal mains. That mix matters more than it first appears, because different materials fail in fundamentally different ways. Older metal mains tend to announce themselves — when they fail, they typically burst, and the water surfaces where someone can see it. Plastic pipe is quieter. It can weep for months underground before any visible sign appears at the surface at all. A physical inspection programme built around visible failure will systematically miss the failure mode that a third of the network is most prone to.
It's worth being precise about what's real today versus what's genuinely speculative. The National Underground Asset Register (NUAR) is a real, government-backed digital map of the UK's buried utility infrastructure, rolled out to reduce the roughly £2.4 billion a year in economic damage caused by accidental strikes during excavation. NUAR is genuinely valuable, and it genuinely exists — but it answers a location question: where is this asset. It does not answer a condition question: what state is this asset actually in.
That second question is where a genuine digital twin vision comes in, and where the honest answer today is that it doesn't fully exist yet for buried pipe at national scale. The idea: combine what's already known about a pipe (material, age, installation record, soil type, historical leak and burst data nearby, pressure readings, acoustic monitoring data where it exists) into a model that produces a probabilistic condition estimate for every metre of pipe in the network, continuously updated as new evidence arrives, rather than waiting for a physical survey that may never happen.
The honest, necessary caveat: any such model produces a confidence-rated estimate, not a fact. An asset with strong surrounding evidence (recent nearby survey data, known material, stable pressure history) deserves a high-confidence estimate. An isolated pipe with none of that context deserves an honestly low-confidence one — flagged as a genuine priority for physical inspection, not silently treated as fine because a model produced a number. This mirrors exactly the same discipline any AI-assisted condition assessment should follow: a proposed grade with a stated confidence level, reviewed by a qualified engineer, never presented as certainty it doesn't have.
Not a national buried-network digital twin on day one — that's a genuinely large undertaking. A realistic first step looks like: pick a bounded, well-instrumented section of network with real acoustic or pressure monitoring data already flowing, build a probabilistic model for that section specifically, and validate it against whatever physical inspection or leak history already exists there. Prove the confidence calibration is honest on a small scale before ever trusting it at a larger one.
An earlier piece on this page laid out the scale of the problem: close to half of the UK water treatment workforce is over 50, and the sector needs an estimated 43,700 additional workers by 2030 just to replace retirees and deliver AMP8. This piece asks a narrower, more practical question: what can AI actually do about it, honestly, and what can it not do at all?
Research widely cited across knowledge-management literature puts the split at roughly 80% of an organisation's real operating knowledge held as tacit, undocumented experience, against only 20% ever formally written down. One widely cited industry study found that around 92% of organisations still fail to consistently capture knowledge from retiring employees, even where a formal exit interview process already exists. The reason is structural: an exit interview is one conversation, on one day, capturing only what the departing expert happens to remember to mention in the room. It was never built to capture judgement applied in the moment, months or years earlier.
Three genuinely different capabilities, not one single trick. First, mining what already exists in unstructured form — years of work orders, inspection notes, incident reports, and email threads — to surface patterns an engineer would recognise instantly but that have never been pulled together in one place. Second, structured, AI-guided interviewing that goes well beyond a single exit conversation: shorter, repeated sessions held over the final year or two before retirement, designed around techniques from cognitive science specifically to help an expert articulate judgement they've never had to put into words before. Third, building the connections between people, decisions, and outcomes into something genuinely searchable, so a future engineer can ask a real question and find an answer grounded in a specific, real precedent, rather than a generic manual.
This matters enough to state plainly, not softened. A large language model cannot replicate embodied, hands-on skill — the specific feel of a tool, the sound that tells an experienced technician a bearing is failing before an instrument confirms it. That kind of knowledge is transferred through apprenticeship and practice, not conversation, and no amount of interviewing changes that. Structured AI interviewing can realistically surface a meaningful share of what an expert can articulate when asked well — but "what can be articulated" and "everything the expert actually knows" are not the same set, and treating them as identical would be dishonest.
The two problems are the same problem, seen from different angles. An asset health system depends on judgement being captured accurately — not just a grade on a form, but the reasoning behind it, and how much confidence to place in it. A retiring engineer's tacit knowledge about which assets carry more real risk than their paperwork suggests is exactly the kind of judgement an asset health system's data governance group is meant to protect. Solving the ageing-workforce data problem and building a genuinely mature asset health system turn out to require exactly the same discipline: capture the reasoning, not just the number, and be honest about how much confidence to place in each one.
One of the most common challenges in asset management is deciding where to invest limited resources. Historically, organisations often focused on the assets in the worst condition and assumed these should be replaced first. While intuitive, this approach can lead to poor investment decisions — an asset in poor condition is not always the asset that poses the greatest business risk, or delivers the greatest benefit if renewed.
Modern asset management has evolved beyond condition-based thinking. The key question is no longer "which assets are in the worst condition?" Instead, mature organisations ask: "which investments create the greatest value for customers, stakeholders and the business?"
This answers: which assets are most important to the business? Criticality measures how important the asset is to delivering its intended function. Consequence measures what happens if it fails. Value measures the benefit it provides to customers, operations, regulatory performance, resilience and organisational objectives.
| Asset | Criticality | Consequence | Value | Score |
|---|---|---|---|---|
| Strategic reservoir | 5 | 5 | 5 | 125 |
| Workshop heater | 1 | 1 | 1 | 1 |
Both assets may need maintenance, but the reservoir is clearly more important to the organisation. This becomes an Asset Importance Register.
This answers: which assets are deteriorating? Knowing an asset is important doesn't mean it's actually deteriorating — this layer measures physical condition alongside its ability to deliver the required service.
| Asset | Condition | Performance | Score |
|---|---|---|---|
| Pump A | 4 | 4 | 16 |
| Pump B | 2 | 2 | 4 |
Pump A requires greater attention, showing more significant deterioration and poorer performance. This forms an Asset Health Index.
Neither layer should be used in isolation. An asset can be highly critical but still perform well. Equally, an asset can be in poor condition but have little impact on service delivery. The final stage combines both.
Where Performance Efficiency = (Performance + Cost) ÷ 2. This answers the question every investment board actually cares about: where should we spend money first?
It brings together five fundamental dimensions in one score: importance (criticality), business impact (consequence), strategic value, physical health (condition), and operational effectiveness (performance and cost) — moving from a simple risk ranking into a genuine investment decision model.
Despite the site lighting being in far worse physical condition, it provides limited business value and presents minimal risk to service delivery — correctly ranking well below the pump.
Asset Class Strategies, Asset Health Indices, criticality frameworks and portfolio optimisation tools all exist to answer one fundamental question: how can we create the greatest value from our assets? Combining asset importance, asset health and investment priority into one structured framework moves organisations beyond simple risk ranking, toward a genuinely evidence-based asset investment model — a stronger foundation for long-term planning and sustainable service delivery.