The team can prompt, but can it direct?

A portfolio manager at a family office sends you a first draft of the investor update. It came back from Claude in forty seconds. It reads fluently, hits the house style, includes the right metrics. You scan it, approve it, send it on. Two days later, the client flags three errors of emphasis, a missing risk disclosure, and a tone that undersells the quarter. The model was fine. The capability transfer was not.

Most property teams now have access to decent AI tools. The tools work. The problem is rarely the model. The problem is that prompting is not the same as directing, and most teams were never shown how to manage work that comes back in three seconds instead of three days.

Why does AI work for some asset teams but not others?

The difference is not access to better tools. It is whether the team inheriting the system has the capability to direct it. Ethan Mollick at Wharton frames this clearly in his research on AI and management: the bottleneck is not the agent, it is the human’s ability to evaluate, correct, and steer the output. Property teams without that capability get fluent nonsense, or work that looks right but misses the brief.

What is capability transfer in a property engagement?

Capability transfer means the next associate can run the system without me. That includes the brand guidelines, the content templates, the CMS, the image library, the signage spec, and the rhythm of what gets updated when. It also means knowing what the AI can and cannot do, how to review its output, and when to override it. I document this from day one because the handover is not an afterthought.

What does that look like in practice?

For a recent launch, I trained the asset team on the custom CMS, walked them through the content calendar, and ran a session on how to commission new collateral using the templates and tone rules I had built. The AI tools were already in their stack. What they needed was the management layer: how to brief, how to assess, how to maintain consistency when the next leasing suite goes live or the investor report is due.

Why does management capability matter more than the prompt?

Because prompting is the easy part. A decent model will give you fluent, on-brand prose from a half-formed instruction. But it will not know that the planning committee cares more about massing than materials, or that this particular investor hates passive voice, or that the last quarterly update buried the vacancy rate in paragraph nine. A team with management capability knows what to look for, what to fix, and what to kill.

What happens when teams skip the capability transfer?

The system decays. The brand guidelines get ignored because no one remembers why the lockup matters. The CMS fills with off-template pages because the new hire was never shown the structure. The AI gets used for first drafts that no one knows how to assess, so everything gets rewritten from scratch anyway. Six months later, the client calls me back to restore what was already delivered once.

How should a small property team adopt AI in 2026?

Start with one repeatable process where quality is measurable and the stakes are containable. Investor updates, leasing summaries, site progress notes. Use AI to draft, but assign someone with subject knowledge to review and refine. Do not automate sign-off. Build evaluation criteria before you scale the tool. McKinsey’s recent work on generative AI adoption across industries shows the same pattern: early wins come from augmenting skilled judgement, not replacing it.

What does AI literacy mean for an asset manager?

It means knowing what the model is optimising for, which is fluency and coherence, not accuracy or commercial judgement. It means understanding that the output will always need a human with domain knowledge to check it against the actual asset, the actual brief, the actual risk. AI literacy is not about learning to code. It is about learning to manage a new category of collaborator that works fast, never tires, and has no commercial instinct. This is what I mean when I say AI arms teams, it doesn’t replace them.

Why do I build handover into every engagement?

Because the work I do only succeeds if it outlives my involvement. A pitch document delivered three weeks before the meeting is useless if the team cannot adapt it when the client changes the terms. A brand system is useless if the next marketing associate cannot apply it to the hoarding going up on Friday. Capability transfer is not a nice-to-have. It is the difference between a deliverable and a system.

I am not trying to make teams dependent on me. I am trying to make them capable without me. That includes file sources, font licences, content rules, and the management capability to commission, assess, and refine work at the speed AI now makes possible. If your team has the tools but not the capability to direct them, I would love a chat. Send a paragraph on what you are trying to do. I will write back.