How CRE Analysts Will Best Learn the Work That AI Can’t Do

For decades, the path to becoming a competent asset manager or acquisitions principal was unglamorous and repetitive. A junior analyst spent their first two years buried in offering memoranda, lease abstracts and trailing 12-month financials. They made the small, expensive mistakes that taught them what to look for. By year five, they could look at a rent roll and intuitively know which tenants would renew before the leasing team even made the call. 

The work was tedious. That was exactly the point.

That apprenticeship is now in a direct collision with the technology meant to liberate it.

A purpose-built agentic system can read three complex retail leases in under seven minutes. It can produce a structured comparison of uses, rents and escalations while the analyst is still opening the PDF. The same system can run a mixed-use underwrite in ninety minutes. These workflows were once the industry’s primary training ground. Now, they are being automated out of existence.

This part of the AI conversation does not get enough attention in commercial real estate. While we debate which jobs AI will eventually do, we rarely discuss how the next cohort of analysts will learn the jobs AI cannot do.

There are at least five critical roles that AI cannot replace in this industry. These belong to the relationship keepers, the ultimate decision-makers, the handlers of complex edge cases, the strategists who bridge AI to physical operations, and the experts with the intuition to know which option is actually right.

As AI takes on more of the analytical work, relationships become one of the few things you can’t replicate. However, relationships in real estate are built on fluency. To advise a client well, a professional must be well-versed in how deals are structured and where the hidden risks live. 

That confidence comes from thousands of repetitions. Now, those repetitions are being handed to a model. This can create a dangerous knowledge gap.

The burden of solving this gap falls squarely on senior leadership. For years, the sheer volume of manual work forced juniors to learn almost through osmosis. If a senior partner assigns a task to an AI-equipped junior today, they are no longer delegating a task but the junior’s education.

So what does an apprenticeship look like when AI does the reps?

In firms doing this well, the analyst’s first years do not disappear. Instead, they compress. Fifty hours of manual lease abstraction turns into two hours of auditing the AI’s version. However, the real learning only happens if the senior person on the other side of the desk treats the review as a high-stakes interrogation.

Senior staff must move from being almost passive recipients of finished work to active participants in the analytical process. For example, why did the model flag this specific termination right as a low risk? Without this active friction, you simply have a junior employee clicking “approve” on data they do not truly understand. 

Only a fraction of CRE firms say they have achieved their AI goals. The reason being that today most rollouts are treated as software updates rather than training projects. Procurement buys the tool, and operations is told to use it. Nobody is assigned to develop the humans working alongside the machine.

The hiring emphasis needs to shift toward those who can ask sharper questions. Modeling speed matters less than it did five years ago. Agentic AI is changing the economics of every deal team in the country and only the firms that treat AI as a teaching tool will survive the transition.

Arunabh Dastidar is the co-founder and CEO of real estate investment platform Leni.