AI Is Starting to Rewrite the Real Estate Office

A lot of the AI conversation in real estate still sits at the visible end of the business... listing copy, prospecting, appraisals, search, portals, marketing automation. That is understandable, because it is the part people can see. But the more I look at how agencies actually run, and the more of the recent AI research I read, the more it feels like another layer is starting to come into focus... not just tasks, but roles.

Deloitte’s latest enterprise AI work says 84% of companies still have not redesigned jobs around AI. That number stood out... not because real estate should copy the broader corporate world, but because it points to the same gap I keep noticing inside agencies. A lot of offices are experimenting with tools, but the office itself is still organised as if every part of the process needs to be carried manually by a person somewhere in the chain.

That is how most real estate roles were built in the first place. They were not designed from first principles. They grew around friction. Someone had to update the CRM, someone had to chase missing information, someone had to prepare the marketing, someone had to move documents around, someone had to rewrite details that already existed in another system, and someone had to remember what was supposed to happen next. Over time, that becomes an org chart.

Which is why I think the next AI question in real estate is becoming less about what the tool can do, and more about who should still be doing what once the system can carry more of the process.

Take a fairly normal office. A sales administrator might spend part of the day logging notes, re-entering inquiry details, sending standard follow-up, preparing documents, and checking whether each stage has been completed properly. A marketing coordinator might still be taking the same property facts and turning them into listing copy, brochure copy, portal copy, social posts and vendor updates in slightly different forms. A salesperson might still be doing a surprising amount of mechanical follow-up, status checking, chasing and internal coordination around the edges of the actual client work.

Once AI starts handling first drafts, summaries, CRM updates, lead triage, follow-up prompts and document preparation more reliably, those jobs do not disappear in one dramatic moment. But they do start to change shape. Some parts shrink. Some parts move. Some become more about checking, editing, escalating and making judgement calls rather than producing everything from scratch.

That, to me, is where a lot of the larger enterprise research is actually pointing. Bain’s recent Code Red piece is useful on this. Their argument is basically that workflow redesign and workforce redesign have to happen together. McKinsey is saying something very similar in its State of Organizations, that AI is pushing organisations to rethink end-to-end processes rather than just layer software over the top. In real estate terms, that means the issue is not whether the office has tried an AI writing tool. It is whether the office still makes sense structurally once more of the routine coordination work can be done by the system.

You can already see the local signals. REA has been talking about conversational search... but also about how AI is speeding up internal product development and productivity. Domain’s LeadScope is interesting for a similar reason. It is not just a flashy front-end feature. It reaches into the CRM, looks for patterns, and tries to surface likely sellers from existing data. That matters because it shows where the tools are heading. They are moving closer to the actual flow of work.

And once software starts moving closer to the flow of work, the pressure does not just fall on tasks... it falls on the structure built around those tasks.

That is where smaller agencies may have more room to move than they realise. A lean office with clean workflow can often change roles much faster than a larger office with layers of habit, duplicated systems and support functions built around old inefficiencies. If the process is clear, a small team can reassign work quickly. If it is not, AI usually just helps the clutter move faster.

What does remain true, though, is that real estate is still a people business in the parts that matter most. Clients still lean on judgement, timing, reassurance, negotiation, and 'local feel' in ways that software does not replicate particularly well. Where the ground is shifting is elsewhere, in the surrounding office work that has historically needed human handling simply because the systems were weak, disconnected, or too clumsy to carry it properly. That support layer is not disappearing overnight, but it is becoming much easier to reduce, reshape, and fold back into the workflow itself.

So I do not think the right question now is whether AI will replace agents. I still think that is too blunt, and probably aimed at the wrong part of the business.

The more useful question is whether the real estate office is still organised for a world where every part of the process has to be manually carried by someone.

Because if that assumption starts to break... then the office changes with it.

The agencies that get the most out of this will not just be the ones trying more tools. They will be the ones willing to redesign roles around cleaner workflow, better systems, and a clearer division between what should be handled by people, what should be handled by software, and what should never have been repeated manual work in the first place.