Stop chasing the model, start chasing the value
AI continues to move quickly, but not necessarily more clearly. Every week brings a new model, a new claim, or a new reason to wonder whether we are moving too slowly or too fast. Rather than chasing every development, let’s take a more practical view: what should leaders pay attention to now, and what might a sensible next phase of AI adoption look like?
From 'can it?’ to 'is it worth it?'
In “AI and the road to the future”, the INREV Technology Committee made a deliberately measured case. AI should be an enabler, not a disruptor. The core work of allocating capital, managing portfolios and delivering returns is likely to remain largely unchanged—what will change is how quickly we execute it and how far expertise can be scaled across a team. In the last few months, the question of “Can AI actually do this?” has shifted to “Are we achieving real value from this? And is our team getting better, or just faster?”
Successful AI deployment has moved from a technology challenge to more of a leadership one.
Addressing uncertainty
Among surveyors, fund managers, and analysts, often with decades of experience, many are quietly apprehensive about AI. Their concerns are practical, not irrational, and three realities appear. They are:
- Nervous about using AI and the risk of relying on outputs they cannot fully defend.
- Scared not to use it, in case they are left behind.
- Worried about technology moving faster than anyone's ability to keep up.
Pretending these fears can be dissolved by a lunchtime webinar misses the point and could lead to quiet avoidance or reckless over-reliance. Naming the issue is the first act of honest change management, and can help create the space to talk about standards and responsibility without making people feel exposed.
What separates AI users
In a very useful article, David Brooks argues that in the age of AI, people will be differentiated ‘not by how smart they are, but by their relationship to mental effort’.¹ Whilst he raises the risk of cognitive polarisation, this line is worth remembering: ‘when intelligence is plentiful, volition is valuable'.
In plainer language: as knowledge is becoming ‘free’, wisdom is growing more expensive. When any analyst can produce a plausible investment paper in ninety seconds, what remains scarce — and valuable — is the judgement to know whether it is right, and the human skill to explain what it means.
Brooks presents three kinds of AI users emerging in our teams:
- ‘ Productive Passengers’ use AI to move faster, but risk weaker judgement and critical thinking through over-reliance.
- ‘ Reluctant Optimisers’ understand risks, yet drift into convenience under deadline pressure.
- ‘ Mental Marathoners’ treat AI as a challenger or coach, preserving effort, originality and agency.
The uncomfortable implication for leaders is that simply handing out tools is not a strategy and could even be a liability. Our job is to shape how people use them.
Chase the value, not the model
There is still too much noise about models: which one, how large, how new. But the value often sits in a specific, repeatable business problem solved end to end without frustrating manual work and a graveyard of offline spreadsheets. Chasing the model is a hobby. Chasing the value is the job.
The firms that capture value are not necessarily those with the cleverest model; they are the ones that redesigned the workflow and embedded governance early.
Discipline, not novelty, is the real differentiator, and this is what any well-run real estate business already understands.
Scaling without chaos: three pressure points
Three pressure points will decide whether organisations scale AI successfully.
- Measurement matters because "time saved" only counts if it translates into business outcomes, with token, licence and support costs properly understood.
- Governance matters because experimentation can quickly become tool sprawl, duplicated effort and data exposure.
- Change management matters because people will not adopt a new way of working simply because the technology exists.
AI makes these familiar disciplines more urgent, and the real risk is not that a single use case disappoints, but that many small experiments spread without a common view of value and ownership.
Someone must own the AI way of working: not only the IT of it, but the culture, behaviours and judgement that determine whether it improves decisions.
What to do next
We should prioritise practical steps over hype. Firms should choose two or three AI use cases where value can be measured rather than asserted. Tool tiers and human-in-the-loop expectations must be established before scaling creates sprawl. Ownership of AI adoption also shouldn’t sit solely with technology teams, but with business operations and workflows. Investing in AI capability is still important, but successful usage will depend just as much on behaviours and judgement.
In an age where technology paradigm shifts happen increasingly frequently, it’s unsurprising that provocative questions keep arising. These issues fuel the discussions among the INREV Technology Committee and will be explored at the INREV Technology Seminar on 15 September 2026. Join the conversation at the seminar, or share your knowledge with the industry by submitting your own AI case studies to the INREV Case Study Library.
¹ David Brooks, "The Age of Intelligence", The Atlantic, 2026.