Monthly Register — September 2024

September closed without handing anyone a headline. No trade show, no award, no funding print that had to carry the month. That makes it an unusually good month to write down, because what was actually moving in AECO information management last autumn was a set of conditions rather than a run of events, and conditions are the easiest thing in the world to stop noticing. Nobody circulates a bulletin about the weather they work in every day.

Three of those conditions were running at once, and they were running on the same jobs.

What moved

Start with the vocabulary, because that is where it shows first. ISO 19650 had become the language of the job — including on jobs where the common data environment was, in practice, a shared drive with permissions on it. BIM execution plans cited the parts. Exchange information requirements borrowed the phrasing. Suitability codes appeared on title blocks, sometimes correctly. The clause set was genuinely in use. The container discipline that clause set assumes was not, on a large share of live projects, in use in the same way.

You could watch that gap open in a single afternoon. A document controller who can quote the state codes from memory, filing into a folder tree a different consultant invented several years earlier, which nobody since has been brave enough to touch because three parties now have shortcuts pointing at it. The language arrived and the plumbing did not, and by autumn 2024 that was normal rather than remarkable.

Then the sprawl. A mid-size job ran a model platform, a document platform, a field app, a commercial system, and a long trail of email behind all four. Every one of those purchases was rational for the team that made it — the commercial people needed a commercial system, the site needed something that worked on a phone in the rain — and none of them was bought by somebody whose job was the project record as a whole. The assembly then failed on the simplest question anybody asks. A subcontractor rings to check whether the sheet in their hand is still the one to work to. Answering takes four logins, a phone call, and a degree of hoping. Forty systems and no single source of truth was already the diagnosis, not a future risk on somebody's register, and the people who felt it first were whoever got asked to assemble a set for a claim.

Third, in the United Kingdom, the golden thread had moved out of conference vocabulary and into a record duty. The Building Safety Act had put higher-risk buildings under a regulator that expected an identifiable, available, change-controlled set of information — not a folder named "golden thread", filled late and handed over at practical completion with everyone's fingers crossed. What teams were still working out, and most were nowhere near finished working out, was what the duty meant for the issued set on an ordinary Tuesday in the middle of construction, rather than for the handover pack at the end of it. The instinct on most jobs was to treat it as a deliverable and give it to whoever had least on. That instinct is understandable and it is the wrong shape, because a duty about availability is a duty about how you work all the way through, not a folder you populate at the end.

What changed in the standards

Nothing was published that month. The working texts most BEPs still pointed at were the ISO 19650 parts already in circulation — information management through delivery, plus the adjacent parts for assets and security where a client had asked for them. Those were not new in September. They were in use in September, which is a different fact and a more useful one, because a standard in use gets tested in ways a standard in draft never is. Every awkward edge in the container model was being found by somebody that month, usually at half past five on a Friday.

The standards-adjacent argument that actually reached document control in Britain was the Building Safety Act's information duties and the golden thread: what must be kept, who may change it, and whether the record is available when the regulator or the principal accountable person asks for it. That is not a new ISO number. It is a duty that assumes a controlled record, and a good many organisations carrying it did not have one.

Elsewhere, permit-stage BIM and digital-submission rules were already changing what "published" meant at the gate. When a jurisdiction tells you that the model, or the information container, forms part of the permit application, the definition of a finished issue changes for everybody on the appointment — including the people who never open the model and had assumed this was somebody else's problem.

One thing to watch

Watch, through the autumn, whether those three start turning up in the same argument. ISO 19650 language without the container discipline. Sprawl treated as normal. A UK duty that quietly assumes a single, change-controlled record. They read as three separate topics. In my experience they are one information-management problem in three dialects, and the tell is that the same handful of people keep getting pulled into all three meetings.

Two things worth doing about that, rather than nodding at it. On your own job, ask who can name the current issued set without opening five systems, and time them; if the answer is one person and they are on leave next week, that is your finding, and it is a resourcing problem before it is a software problem. And when a client cites ISO 19650 in an appointment, ask at appointment stage which parts they mean and what they expect the states to do. That conversation costs an hour before anyone has built a folder tree, and considerably more afterwards.