Will AI Replace Real Estate Agents in New Zealand?

We have spent money on AI this year that has returned nothing yet. I would spend it again tomorrow.

That is an odd thing to put in writing, so let me explain why I think it is the right call.

Will AI replace real estate agents?

No. Not in this market, and not at the price points we work in.

A $5M+ sale runs on judgement under pressure, on reading what a buyer has not said, and on a vendor trusting one person enough to take their advice when it is not what they wanted to hear. None of that is a language problem, and language is what these systems are good at.

The question is more useful turned around. AI will not replace the agent. It will change what one agent can reasonably cover in a week, and that is a different kind of pressure.

What does AI actually change in a real estate business?

The repetitive middle. Not the relationship, and not the advice.

The work that disappears first is the work nobody was proud of anyway. Drafting the third version of a schedule. Summarising a long report into something a vendor will actually read. Pulling six months of campaign data into a shape you can look at. Rewriting the same email for the eleventh time.

That is not glamorous work, but it is hours, and hours are what an agent runs out of.

The agents I see doing well with it are not using it to sound smarter. They are using it to get ninety minutes back and putting that ninety minutes into conversations.

Why is "seamless" the wrong word?

Because nothing about bringing AI into a working business is seamless, and I think the people selling it that way are doing real damage.

Here is what it actually looks like. You pick a tool. It does not do what the demonstration suggested. You rebuild a process around it and the process breaks somewhere you did not expect. Someone on the team hates it. You spend money this quarter for a return that shows up, if it shows up, a year from now.

That is the honest version. It is messy, it is slower than promised, and it costs before it pays.

I say that as someone who is genuinely enthusiastic about this. The enthusiasm is not the problem. Pretending the transition is painless is the problem, because people then abandon it at the first friction and conclude the technology does not work.

What happens to an agent working on their own?

They fall behind, and I do not say that as criticism.

An individual agent carrying their own admin, their own marketing, their own database and their own pipeline has no spare capacity to run a six-month experiment that might fail. That is not a character flaw. It is maths.

What a business can do is absorb that cost centrally. We can test something badly, learn from it, and only hand the team the version that works. The individual agent gets the outcome without paying the tuition.

This is the part of the brand conversation that has changed most in the past two years. It used to be about reach and reputation. Increasingly it is also about who is carrying the cost of figuring this out on your behalf.

Why is waiting the expensive option?

Because the cost of starting does not fall, it just moves.

The argument for waiting is reasonable on the surface. The tools are improving fast, so why buy into this year's version when next year's will be better and cheaper.

The flaw is that the tool was never the hard part. The hard part is the process change around it, the team habits, knowing which tasks are worth handing over and which are not. That takes months regardless of how good the software is, and none of it gets easier by waiting.

I would rather be six months into the messy part than starting it next January.

Where I think this lands

The agents who do well through this will not be the most technical.

They will be the ones who worked out early which parts of the job were actually theirs, and handed over everything else without sentiment.

The job itself is not under threat. Sitting with a vendor who has just had a difficult month, and telling them the truth about their price, is not a task anyone is automating.

What is under threat is the version of the job where a capable person spends half their week on work that no longer needs a person.

That is worth getting ahead of. Not because the technology is exciting, though I think it is, but because the gap between the agents who adapted and the agents who waited is going to be visible well before anyone announces it.