AI Is Starting to Reprice Real Estate Work.

Over the weekend, I came across a strong piece from Phil Fersht on LinkedIn, titled "Companies are not firing people. They are quietly closing the front door on the next generation of knowledge workers". His article was built around Anthropic's new report, Labor market impacts of AI: A new measure and early evidence, and it struck me because it put some structure around something I have been thinking about for quite a while.

The AI discussion in real estate still tends to start with whether agents themselves are under threat. I can understand that, but I do not think that is where the first real change is showing up. To me, the more immediate pressure is building in the work around the agent, the drafting, follow-up, admin, coordination and all the system gaps that agencies have been carrying for years.

That is one reason the Anthropic research landed for me. It is not really describing some sudden white-collar collapse. It is describing a slower shift, where certain kinds of work become easier to assist, automate or compress, and where hiring and role design start changing around the edges before the more dramatic headlines appear. That feels much closer to what I think real estate is heading into.

Real estate has always had a lot of manual process built into it. Listing copy gets rewritten when it could have been written once and reused properly. Marketing tasks get spread across texts, calls and emails. Follow-up depends too much on memory. Data gets entered more than once. Admin ends up filling the gaps between systems that do not talk to each other properly. None of that is new. What is changing is that there are now much better ways of dealing with a lot of it.

For years the industry managed that friction by adding people. If the business became more complicated, more coordination was added, more handoffs, more support, and usually more software layered over the top of the same old problems. That worked while labour was carrying the weight, but it becomes harder to justify once AI starts reducing the amount of work needed around drafting, updating, summarising, follow-up and low-level coordination.

That does not mean the human side of the job suddenly stops mattering. In real estate, trust still matters, judgement still matters, negotiation still matters, timing still matters, local knowledge still matters, and reading a client properly still matters. What changes is the amount of labour needed to support everything around that layer.

Part of that is obvious enough. Campaign drafts, first-pass emails, CRM updates, call summaries, document preparation, follow-up sequences and vendor reporting all sit in the firing line. Part of it is less obvious, moving information between systems, chasing up the next step, filling the day with low-level coordination, and relying on someone to remember what should happen next because the workflow itself has never really been sorted out.

As that layer starts to shrink, the economics of the agency start moving as well. A solo operator can handle more without feeling constantly buried. A small team can produce more consistent work without the same jump in overhead. A larger office that has been relying on people to hold together weak process starts to feel heavier than it used to.

Workflow architecture matters more than tools

That is why I keep coming back to workflow architecture, or more simply, how the work is actually set up.

A lot of the conversation still gets stuck on tools. Which model is best. Which app saves time. Which platform writes the best copy. Those things matter, but they are not the main issue. If the workflow is messy, AI usually gets dropped into the middle of that mess and the result is average at best. The bigger gains come when the business has already done the harder thinking about how information should move, where tasks should sit, what should be automated, what should be templated, what should be outsourced and what should stop happening altogether.

Those are business questions before they are technology questions.

The early signals are already visible

The strongest local signals are already coming from REA and Domain. REA Group has rolled out AI features around valuation interpretation and natural-language search. Domain has been using AI internally to surface listing and property data more quickly. If you look outside Australia, the same direction is obvious there as well. Zillow is talking about AI inside CRM, call summaries and workflow support for agents. Compass, eXp and others are pushing further into contact management, transaction support and business insights. The broad direction is not hard to see. AI is moving into the systems layer of real estate, which is also where a lot of inefficiency has been sitting for years.

That matters even more for smaller operators. Big businesses have money, brand and reach, but they also carry legacy process, messy software stacks, internal approvals and habits that are hard to unwind. Independent agents and small teams are often in a better position to simplify. They can redesign faster, make cleaner decisions and build a sharper operating model without dragging old structures along behind them.

That does not mean anyone should become careless with these tools. Real estate is public-facing, heavily marketed and full of compliance risk. Australia is already seeing scrutiny around AI use in listings and visual presentation. Agencies that use these tools lazily are going to create a new set of problems for themselves. The people who get the most out of this shift will be the disciplined operators, the ones who combine AI with checking, disclosure, strong process and good judgement, while keeping humans in the loop wherever that is practical and possible.

Efficiency pressure arrives before the headline shift

This is why I see the early impact of AI in real estate as an efficiency story first. The pressure is building around work that has been tolerated for years because there were not many alternatives. Slow admin, duplicated effort, weak handoffs between systems and support roles built around avoidable friction all become much more obvious once a cleaner way of operating starts to emerge.

In a tighter market, this starts to matter quite quickly. Margins tighten, responsiveness matters more, and clients notice the difference between a business that feels organised and one that feels cluttered. Agencies carrying too much operational drag will feel that pressure earlier than they think. The ones that deal with it sooner will simply run better.

When I think about what AI means for real estate, I keep coming back to workflow. The real estate industry has carried too much repeated admin, too many clunky handoffs and too many systems that were never set up properly. AI puts all of that under pressure, and to me that is where the first serious change is already starting to land.