If you've ever spent a Friday afternoon cross-referencing a site against flood zones, conservation areas, Green Belt boundaries, and three different council policy documents, you'll already know the honest answer to "how long does planning due diligence take?" is: longer than anyone would like.
It's not that the work is complicated, exactly. It's that it's slow, repetitive, and spread across a dozen different sources that rarely agree with each other on formatting, let alone substance. And with the new NPPF landing on 17th August 2026 - the first comprehensive rewrite since 2012, according to Landmark Chambers' analysis - a fair few of the policy assumptions land teams have relied on for years now need re-checking from scratch.
So it's a reasonable moment to ask: can AI actually take a meaningful chunk out of that timeline? Or is "AI-powered due diligence" mostly marketing sitting on top of the same manual process?
Here's our honest, unspun answer.
Before you can even think about whether AI helps, it's worth being honest about what "due diligence" covers, because it's a lot more than checking if a site has planning permission.
At the site-triage stage, you're typically trying to establish:
Traditionally, that means opening the local plan, the planning portal, the Land Registry, an ordnance survey layer, and probably a council's own interactive constraints map, none of which were designed to talk to each other. It's not a glamorous process, but it's the process that decides whether a site is worth pursuing at all.
Statutory decision periods in England are set at 8 weeks for most applications and 13 weeks for major development, under the government's own guidance on determining a planning application. In practice, that clock only starts once an application has been validated, which itself is only possible once due diligence, research, and drafting are already finished. Anecdotally, anyone who's worked a live pipeline will tell you the same thing: extensions of time and agreed longer timelines are common rather than exceptional, particularly on major schemes.
We're currently pulling together our own research into exactly how long applications are really taking across England, so watch this space.
What we can say with confidence is this: none of that statutory clock accounts for the due diligence a developer does before an application is ever submitted, which is exactly the stage where speed matters most, because it's where you decide whether a site is worth pursuing at all. Shaving weeks off your own research process, before you're even in the council's queue, makes a genuine difference to how many sites you can realistically assess.
Yes, with a fairly important caveat.
Generic AI tools like ChatGPT can absolutely summarise planning policy or draft you a paragraph about flood risk. What they typically can't do reliably is tell you where that information came from, or guarantee it's still accurate once policy has moved on (as it just has, quite significantly). That's a real problem in a discipline where the wrong assumption about a site's status can cost you weeks, or a deal.
This is where the distinction between "an AI chatbot bolted onto a planning question" and "AI built specifically for planning triage" starts to matter. We've written before about how AI actually identifies planning constraints on a site and, separately, what to check before trusting any AI-powered land sourcing tool - the short version is that AI doesn't invent constraint data out of nowhere, it cross-references it against a dataset, and the quality of that dataset (and whether you can actually see it) is what separates a useful tool from a confident-sounding guess.
Done properly, AI can radically reduce the double-checking tax of early-stage triage: instead of manually confirming every constraint yourself, you're verifying an answer that's already been generated and sourced for you. That's a meaningfully different, and faster, starting point.
To be clear about what we're not claiming: AI isn't going to complete your appraisal for you, and it shouldn't be making the final call on a multi-million-pound acquisition. Planning due diligence at the sharp end still needs a human who understands the nuance of a local committee, the politics of a particular authority, or the small print of a Section 106 obligation.
What AI is genuinely good at is the unglamorous, high-volume, early-stage work: giving you a fast, sourced first read on a site so you can decide whether it's worth a planner's time at all. It's an accelerant for the initial sift, not a substitute for professional judgement further down the line.
If you're weighing up whether an AI tool is actually going to save you time, or just move the work around, it's worth asking:
If the answer to the first two is "no," you're not doing due diligence any faster - you're just adding an extra step to double-check the AI's homework.
AI can speed up planning due diligence, but only the part of it that's genuinely suited to automation: fast, sourced, first-pass triage across constraints and policy. It's a head start, not a shortcut past the parts of the job that still need a person behind them.
If you want to see what that looks like in practice, LandInsight's AI Assistant triages sites with citations linked straight back to the planning record behind every answer, so you're never just taking its word for it. You can read more about how it works in our product update introducing the AI Assistant, or head to our resource centre for more on navigating the new NPPF.