AI is moving so quickly that writing an article about it feels slightly dangerous. By the time you finish reading this, there is every chance the tool everyone was raving about last Tuesday has been overtaken by something launched on Thursday by a company nobody had heard of on Wednesday.

Welcome to artificial intelligence.

For property developers, though, the interesting question is not really which AI platform is currently winning the popularity contest. The much bigger question is: how can you use AI to gain an advantage when finding and analysing property deals?

Stop Looking for the Magic Property Tool

One of the most common questions I hear is, “Which AI tool are you using to find deals?”

The assumption is that somewhere out there exists a magical subscription platform where you press a button marked Find Me Profitable Development Sites, make a cup of coffee and return five minutes later to discover a list of perfect opportunities complete with zoning, overlays, comparables, yield analysis and vendors already emotionally prepared to accept your offer.

Sadly, no.

There are plenty of excellent property platforms, and many now incorporate varying amounts of AI into their systems. But there is a commercial reality worth remembering: their business model relies on you continuing to subscribe.

Imagine a tool that genuinely analysed every property in your target area, identified exactly which ones suited your strategy, ranked them, completed the research and handed you the definitive answer could potentially be a terrible business model. You might subscribe once, run the analysis, collect everything you needed and disappear into the sunset without even completing the cancellation survey.

Wonderful for you. Potentially less wonderful for the company selling the subscription.

That doesn't mean property software companies are sitting in dark rooms deliberately hiding the best development sites from you while laughing maniacally. It simply means the perfect “do absolutely everything for me” AI platform may never be the most commercially attractive product for somebody else to build.

Which brings us to the more interesting alternative: build the capability yourself.

Invest in Your Own AI Capability

You do not need to become a programmer, learn seventeen coding languages or suddenly start wearing a hoodie and referring to yourself as a tech founder. What you do need to understand is how to make AI genuinely useful in the way you already work as a property developer.

Used properly, AI can assist with site analysis, market research, planning investigations, council research, comparable sales, development feasibility inputs and all sorts of repetitive tasks that previously involved opening 27 browser tabs and eventually forgetting which one contained the information you were actually looking for.

But the competitive advantage doesn't necessarily come from the AI platform itself. It comes from what you teach it to do.

Two developers can use exactly the same AI model and get completely different outcomes because one has built a clear, detailed process around their strategy while the other has typed, “Is this a good development site?” into a chat box and hoped for divine intervention. AI is clever, but it's not psychic.

Start With Prompting

The foundation is learning how to give AI clear instructions.

Call it prompt engineering if you want to sound impressive at dinner parties, but essentially it means explaining precisely what you want, what information the AI should consider, what rules it should follow and how you want the answer presented.

A vague prompt usually produces a vague answer. A carefully constructed prompt based on your actual development methodology can produce something considerably more useful.

For example, rather than asking AI to simply “research this property”, you can instruct it to investigate zoning, overlays, site dimensions, orientation, planning constraints, local requirements and every other factor that matters to your particular strategy.

Once you have developed a prompt that consistently works, you can turn it into a repeatable process. That is where AI becomes genuinely powerful.

Build Specialists, Not One Giant Robot

There is a temptation to build one enormous AI instruction that asks it to research the site, interpret planning controls, analyse the market, calculate potential yield, assess comparable sales, identify risks and possibly make lunch.

More is always better, right? Wrong.

It sounds efficient, but in practice it can become a little like putting one person in charge of town planning, quantity surveying, sales research, design, finance and making the coffee. Something is eventually going to suffer.

AI models work with something called a context window — effectively the amount of information they can actively process at once. Stuff too much into one task and performance can deteriorate.

A better approach is often to create several highly specialised AI workflows. Think of them as members of your development team.

You might have one specialist whose sole job is site research, gathering zoning, overlays, lot dimensions and basic property data. Another might focus only on council requirements and planning rules relevant to your strategy, while another could analyse market statistics and comparable sales. You could then have another process that takes the information produced by those specialists and turns it into a summary, yield analysis or development report.

Each one becomes extremely good at one narrowly defined task rather than being moderately useful at twenty.

Eventually, these specialist processes can be connected so that you speak to one central AI process, which then coordinates the specialists behind the scenes. The site researcher gathers the basic property information, passes it to the planning researcher, the planning researcher hands its findings to the market analysis process, and eventually everything reaches the reporting tool that produces the final summary.

Essentially, you have built yourself a small virtual development team, with the added bonus that nobody complains about Monday mornings and nobody sends an invoice for attending a meeting to discuss arranging another meeting.

Your Knowledge Is Still the Advantage

There is an important catch. AI does not replace development knowledge. If anything, it makes knowledge more valuable.

You need to know what information matters, what questions should be asked and when an answer looks suspicious. An AI system trained around a poor development strategy simply allows you to make bad decisions significantly faster.

The real competitive advantage comes from combining your property knowledge with AI. If you understand development and then build AI processes around the way you assess sites, those processes begin to reflect your own experience, criteria, assumptions and strategy, which makes them far more valuable than simply subscribing to the same software everyone else is using.

Forget the AI Horse Race

As for which AI platform you should choose, that is probably the least important part of the conversation.

Today, you might prefer ChatGPT. Someone else might swear by Claude. Six months from now, something entirely different may be dominating the conversation.

That is not the important part.

The major AI companies will continue leapfrogging each other, adding features and improving their models. Trying to permanently choose the “best” AI platform is probably about as productive as trying to choose the fastest horse while the race is still running.

The smarter approach is to build the skills that transfer regardless of which platform happens to be leading this month. Learn how to prompt properly, build repeatable workflows, create specialist tools and keep refining them using your own property knowledge.

Because ultimately, the competitive advantage is not the AI.

It is what you know how to make AI do. And unlike another monthly subscription, that is an asset you actually get to keep.