The short answer:

According to research by Places for People and the University of Cambridge, 750,000 construction workers are expected to retire by 2036. That's over a third of the current workforce and the knowledge they carry with them isn't in any AI training dataset. As the industry rushes towards digital transformation, the real risk isn't being left behind by technology. It's discarding decades of hard-won experience before we've figured out how to capture it and lose it forever.

Blog: AI in construction: what happens if we lose the knowledge and can it replace it?

A few weeks ago, I took part in a fireside chat at The Women In Construction and Engineering Awards Summit where someone in the audience suggested we should let decades of construction industry knowledge retire with the people who hold it and let AI start from scratch. I've been sitting with that ever since, because it just doesn't feel right.

And I think it should make all of us a bit uncomfortable, both as technical professionals and marketers and communications leads who are facing exactly the same challenge.

What knowledge is actually at risk as AI adoption grows in construction?

The scale of what's coming is significant. Research by Places for People and the University of Cambridge puts it plainly: 750,000 construction workers are expected to retire by 2036, with 35% of the current workforce already over 50 and only 20% under 30.

What walks out with those people isn't just technical skill. It's the decisions that went wrong and why. The projects that overran because of something nobody saw coming. The client relationships that took a decade to build and aren’t captured anywhere. The instinct that only comes from having lived through a crisis cycle or two.

None of that is in a training dataset and therefore can’t be replicated.

AI works with what's been documented, measured and fed into the model. Large language models have been trained on existing data, they spot patterns and replicate what already exists. The gap where genuinely new thinking happens, where you identify what the tool has missed, is where experience lives. And right now, that gap is at serious risk of being written off as unnecessary.

This isn't both a construction industry and marketing problem

The knowledge retention problem is just applicable to built environment marketing as it is to the wider construction industry.

The numbers on the marketing side are stark. According to Gartner's 2025 CMO Spend Survey, 39% of CMOs are actively reducing headcount and cutting agency budgets, in part because of AI. A separate Gartner survey found 26% of marketing leaders were already planning staff reductions specifically because of generative AI. Meanwhile, LinkedIn's 2025 B2B Marketing Benchmark Survey reports that 95% of B2B marketers are now using AI at least weekly, and 65% are using it daily.

AI tools are everywhere. But the question isn't whether people are using AI, it's whether they're using it well enough to avoid alienating their clients and audiences in the process.

Audiences are getting savvier. They're increasingly calling out what some are bluntly labelling 'AI slop', generic, pattern-driven content that sounds plausible but carries no real authority or sector understanding. In an industry built on trust and long sales cycles, that's a reputational risk that compounds quietly.

Senior marketers in this sector carry years of accumulated understanding: how the industry actually operates, which messages land with which audiences, how to navigate a complex and relationship-driven sales cycle, what genuinely differentiates one practice from another. That knowledge has been built through hard fought experience which can’t be achieved through vague prompting alone.

I've explored this in more depth in a recent piece for The Executive Magazine, Is marketing experience being devalued in the age of AI? [LINK] if you want to dig into the marketing-specific argument further.

Why experience matters even more when everyone has the same tools

If everyone has access to the same AI tools, what does a business actually compete on? The answer is judgement which is built through experience.

Research from the LSE found that stronger performers gained significantly more from AI support than weaker ones, because they could refine and challenge outputs rather than accept them at face value. The same principle applies directly to construction and marketing: without first-hand experience and working knowledge of the task at hand, it's all too easy to accept the first thing AI produces as good enough.

A study by Harvard Business School and the University of California at Berkeley reinforces this. They researched business owners using AI and found that people with stronger existing skills achieved better outcomes, because they had the judgement to identify which AI suggestions were actually worth acting on.

Knowing what good looks like has never been more important. And that's precisely what experience gives you both in construction project delivery and in marketing.

What does the data gap in construction AI actually look like?

Historical data reflects historical inequities. If that's what's being fed into AI tools, the industry isn't transforming, it's automating the same structural problems in a shinier format.

This applies directly to marketing. If the content, messaging and campaigns that inform AI outputs have historically underrepresented certain communities, clients or project types, the tools will reflect that absence. Community consultation is still too often treated as a tick box rather than a genuine source of insight. Procurement processes built for a pre-digital world add months to projects before meaningful work begins.

The same principle holds for workforce planning tools that don't account for who's actually on your team, and marketing platforms trained on data that doesn't reflect the full diversity of the sector's audiences. Better inputs produce better outputs. The reverse is equally true.

What should construction firms and their marketing teams do right now?

This article isn't an argument against AI in construction but it is an argument for bringing the right people into the room when it gets implemented. A few things that actually matter:

Be in the room when the brief gets written – rather than when the tools get rolled out. Whether you're a project director or a head of marketing, if you're not shaping what the digital transformation is trying to achieve, you'll end up implementing someone else's decisions.
Document what you know. If expertise only lives in people's heads, it won't survive a transformation process. I see this constantly in repositioning work with practices, 20 years of knowledge that nobody has ever made visible, and it's exactly what's at risk when senior people retire or move on.
For marketing teams specifically: capture your sector knowledge now. The messaging that works, the audience insights that took years to develop, the positioning that genuinely differentiates your practice. That institutional marketing knowledge needs to be written down before it walks out the door, it’s business critical information.
Frame knowledge retention as risk mitigation as it will ultimately protect the business. Tools built on flawed assumptions create expensive problems downstream whether that's a project that overruns or a marketing campaign that misses the mark with clients who can spot the difference.
Ask who gets to decide where technology steps back. There are areas — sensitive community projects, long-term client relationships, brand positioning those where human judgement should stay primary. At the moment, that call is largely being made by people with a financial interest in more technology – so why would they say no?

What's worth keeping as construction and marketing teams adopt AI?

Not everything needs to change. Some things genuinely need protecting as this industry evolves:

First-principles thinking. No tool replaces the ability to go back to basics and ask why, in project delivery or in a marketing strategy.
Relationships. Construction runs on trust. So does good built environment marketing. Neither gets automated.
Sector knowledge, the embedded understanding of how this industry actually works, who the real decision-makers are, and what genuinely matters to clients at different stages of a project.
Professional judgement built over years. In construction and in marketing, experience is exactly what should be guiding AI implementation rather than being replaced by it.

The professionals and marketers with the most contextual knowledge are often the quietest in the AI conversation. That needs to change.

Frequently asked questions: AI in construction

Is AI replacing construction professionals and marketers?

AI isn’t replacing construction but is changing what's valued. AI in construction is shifting the premium from information-gathering to judgement, context and interpretation. For marketers in the sector, the risk isn't replacement; it's that experienced practitioners disengage from the AI strategy conversation and let tools get shaped without their input.

What is knowledge retention in construction and why does it matter for AI?

Knowledge retention in construction refers to capturing the expertise, decision-making experience and sector understanding that experienced practitioners carry, much of which has never been formally documented. With 750,000 workers expected to retire by 2036 (Places for People / University of Cambridge research), that undocumented knowledge is at serious risk of being lost entirely, leaving AI tools trained on incomplete or historically flawed data.

How should construction firms approach AI adoption?

Start by treating experienced practitioners including your marketing and communications leads, as a core part of the implementation process, not just end users. Document institutional knowledge before it retires with the people who hold it. And challenge the assumption that faster adoption is always better. Tools built on flawed inputs create expensive problems downstream.

What are the risks of AI in construction if knowledge isn't captured first?

The main risks are repeating avoidable mistakes, automating historical inequities into new processes, and producing marketing and communications content that lacks the sector-specific insight clients actually respond to. The gap between planning to implement AI and doing it well is where most of the damage happens.

How does AI affect marketing in the built environment sector?

AI is changing what construction marketing looks like but it can't replicate the sector knowledge that makes built environment marketing actually work. Understanding a long sales cycle, positioning a practice authentically, building trust with technically sophisticated clients that comes from experience, not a prompt. With 39% of CMOs already cutting marketing headcount (Gartner, 2025), the pressure to rely on AI tools uncritically is real. The risk is ending up with content that looks and sounds like everyone else's.

Are you seeing this in your own organisation? Is the knowledge retention question being taken seriously in project delivery and in marketing? Drop a comment below or get in touch — I'd genuinely like to hear what you're seeing on the ground.