There is a number that should stop every director in UK construction in their tracks.
Research by the Get It Right Initiative found that the measured direct cost of avoidable error runs at roughly 5% of project value — around £5 billion a year across the sector. That alone is higher than the industry’s average profit margin of about 3%. Once you fold in unmeasured process waste, latent defects and indirect costs, the estimates climb to somewhere between 10% and 25% of project cost: £10–25 billion every year.
Read that again. The industry spends more on getting it wrong than it earns on getting it built.
We are pleased to confirm that Surtori is now a member of the Get It Right Initiative (GIRI),and joining its technology working group — and we want to be clear about why we joined, and what we intend to contribute.
What GIRI actually is
GIRI is not a marketing badge. It launched in 2017 out of discussions at the Institution of Civil Engineers’ Best Practice Panel, backed by the CITB, with a single, unfashionably narrow aim: eliminate avoidable error in construction.
It is a not-for-profit membership organisation with a deliberately multi-disciplinary membership — clients, consultants, contractors, regulators, educators, insurers, professional institutions and trade bodies. Its working groups cover design, metrics, skills and training, technology, insurance and the Building Safety Act. It publishes research, a Design Guide, and three custom-built training schemes aimed squarely at the root causes of error rather than its symptoms.
Crucially, GIRI’s own research located a significant proportion of error not on site, but upstream — in project formation and design. Errors are made in rooms with laptops and paid for in rooms with concrete.
That framing is the reason we’re here.
Why a digital and AI consultancy belongs in a room about error
There is a reasonable question to ask: what is an AI consultancy doing inside an organisation founded by civil engineers to stop things going wrong on site?
The honest answer is that most error is an information problem before it is anything else.
Across our work on programmes, the same pattern recurs. The defect on site is the visible end of a chain that started somewhere much duller:
- A design assumption that was never written down anywhere findable.
- An RFI answered in an email thread that four people needed and one person saw.
- A specification change that propagated to the drawings but not the method statement.
- A lessons-learned report that was diligently produced, PDF’d, and buried.
- A commercial assumption in the cost plan that quietly stopped being true in month three.
None of that is a competence failure. It is a system failure — an information system operating well beyond the working memory of the humans inside it. And this is exactly the class of problem where applied AI earns its keep: capturing knowledge at the point it is created, surfacing it at the point of decision, and checking consistency across documents no human has the hours to reconcile.
That is Surtori’s whole thesis, and it maps almost exactly onto GIRI’s diagnosis.
What we’re bringing
We joined GIRI to contribute, not to observe. Three areas in particular:
1. Technology that reduces error rather than digitising it. This is why the technology working group is where we’ve chosen to put our time. Construction has spent fifteen years turning paper processes into digital processes and calling it transformation. Frequently, all we’ve done is move the error somewhere with better dashboards. The technology group is one of the few places in the industry asking the harder question — does this tool measurably reduce error, or does it just record it faster? We want to be part of answering that with evidence, and to keep the AI conversation inside it honest.
2. The operating model, not the pilot. Through our Fusion Framework™ and the Digital & AI Operating Model Review™, we spend our time at the unglamorous end: how decisions actually get made, where information actually goes, and why good tools fail inside bad operating models. Error elimination is an operating model question. So is AI adoption. They are the same conversation, and most organisations are having neither.
3. Finding the error before you fund the fix. Our Digital & AI Opportunities™ work exists to answer a question the industry usually skips: where, specifically, is value leaking, and what is it worth? That means tracing error to its origin — design, procurement, commercial control, handover — sizing it honestly, and only then asking what technology has any business being applied to it. It is a diagnostic discipline before it is a technology one, and it sits very close to the work GIRI’s metrics group is already doing.
Get it right, from the start
Twenty-one percent of project turnover, lost to avoidable error. That is not a rounding error, it’s the margin, the bonus pool, the training budget and the innovation fund, several times over.
Surtori joined GIRI because we think the tools now exist to make a real dent in that number — and because we’d rather test that belief in a room full of people who will challenge it than in a brochure.
If you’re a client, contractor or consultant thinking about where AI genuinely reduces error in your delivery model — rather than where it demos well — we’d like to talk.
Surtori is an AI and digital transformation consultancy for construction and the built environment. Find out more at surtori.com, or learn about GIRI at getitright.uk.com.