AI is already paying off for consultants, but human oversight is becoming the differentiator

Another week, another existential crisis. Breakthroughs in artificial intelligence could either save humanity or destroy it, depending on who you listen to. The truth is that no one has yet worked out how to regulate a technology advancing at extraordinary speed. Concerns have steadily grown and are now being expressed in alarming terms: one AI safety expert has warned of a 10% chance that it could kill all humans within a decade. That certainly caught people’s attention.

For most people in business, however, the immediate questions are considerably less existential. What is worth paying for? Will the investment improve productivity or services? And how can businesses make sure the technology is being used responsibly?

Responsibility for the future of humanity is too big for any individual business to take on. In principle, governments can and should use their regulatory powers to slow the development of powerful systems with the potential to outsmart us. But the global race to develop AI, particularly amid competition between the US and China, makes co-ordinated restraint difficult to imagine.

If you are not at the cutting edge of the technology or working in a regulatory role, you are probably more concerned with working out what AI can actually do for your business. That is certainly a major preoccupation highlighted by this year’s Top 150 Consultants survey, in which 87% of firms say they have increased their AI spending. Digging deeper into the responses, it is also clear that many are already able to demonstrate measurable results.

Our survey asked firms how they measure the success of their AI investment, revealing a wide range of metrics tailored to individual business needs. Read enough of the responses, though, and a rough consensus emerges. Firms that have embedded AI tools into their daily work are routinely saving significant amounts of time on administrative tasks such as bids, meeting minutes and document reviews.

Companies may be using different tools and processes, but many report similar productivity gains of between 15% and 30%

Companies may be using different tools and processes, but many report similar productivity gains of between 15% and 30%, with a few reporting figures closer to 50%. If there is a proven return on AI investment in the sector today, the clearest one is reducing the time and cost of administrative work.

It is worth putting some names to those numbers. RED Engineering says it has cut report preparation time by 25%. Bellrock says it has reduced a 40-hour weekly task to two hours. WSP says its work embedding AI into Network Rail’s asset management platform has unlocked an estimated six person-years of engineering capacity a year.

Taken together, these are striking results. They suggest that AI’s near-term business case is no longer theoretical: construction consultancy firms are already able to measure the time and capacity the technology can release.

The more interesting question is what firms do with that capacity. The real prize is not simply doing the same work faster, but using the time saved to create something new or better for the business.

One impressive example comes from Turner & Townsend, which has deployed retrieval-augmented generation (RAG), allowing AI tools to retrieve answers from the firm’s own verified documents. This gives teams better access to programme standards and requirements. Beyond improved efficiency, the consultant says the system has led to better design and BIM quality and better decision-making.

A second consensus is emerging from this year’s responses: the importance of governance and human oversight. Several firms describe building proper processes around AI use, including formal policies, risk registers, boards to test ideas before they are scaled up and technical guardrails to prevent tools straying into areas requiring professional judgement.

The evidence from our survey suggests this is more than hype: there are tangible benefits to be gained, and they extend beyond efficiency alone

Others are candid about why this matters. AI output can look convincing while being wrong, creating a risk that staff treat it as fact without checking. A repeated phrase across the responses is some version of “human in the loop”. While this is primarily seen as a safeguard, a few firms clearly regard it as a potential differentiator in the market. They expect clients will increasingly ask about AI governance when deciding which consultancy to work with.

That points to an important shift. Companies have moved beyond simply asking whether they should invest in AI and are now measuring whether that investment is working. The evidence from our survey suggests this is more than hype: there are tangible benefits to be gained, and they extend beyond efficiency alone.

There are risks too. Being left behind because of underinvestment is one. Failing to invest sufficiently in governance, quality control and human expertise is another.

The firms most likely to pull ahead in AI innovation will be those that can explain to clients and boards not only how AI is improving their products and services, but how they are ensuring human oversight, quality control and professional accountability.

The global debate over AI governance is not going away. But for construction consultants, the immediate question is more practical: can firms prove that AI is making their work better, not just faster – while keeping humans firmly in charge? 

Chloë McCulloch is the editor of Building