You’re not just using property intelligence. You’re shaping it.
Aug 2026
Aug 2026
)
Take the number that tells an insurer how old a roof is. Somebody had to decide that roof age mattered enough to measure in the first place. This meant trial and error as they went about how to read it accurately from aerial imagery. That’s true of nearly everything in property intelligence. From the imagery to the AI scores to the guaranteed pricing, none of it arrived finished. It got built, tested against real properties, and rebuilt when it didn’t deliver.
Most people never see the before. Instead, they only see the output (e.g., an accurate score, a defensible price, an answer you don’t have to double-check). But it’s worth understanding why that work never really stops.
A roof age model built five years ago doesn’t automatically know about roofing materials that didn’t exist five years ago. A fraud pattern an AI model learned to catch this year might not be the same fraud pattern it’ll see next year. The property underneath the data keeps changing, so the tools reading it have to change too, or the accuracy starts to slip.
)
Consider the question you asked your rep last month or the workflow your team wished worked a little differently — all of it lands somewhere. Someone reads it, explores it, and builds toward it. The next tool out of early access could trace back to that moment.
Every time you share your input, you’re helping point the roadmap somewhere. That’s leverage most people using a platform never get.
Standing still is the actual risk
Today’s property intelligence has an expiration date if nobody touches it. Because the property behind that data hasn’t stood still since it was captured. A new building code changes what counts as compliant construction. A catastrophe season redraws which areas carry elevated risk. Eventually, you’re pricing a policy, approving a permit, or budgeting a rebuild against a property that no longer exists.
This problem is already showing up in claims. When a storm rolls through, damage that predates the event often gets attached to the indemnity request. Insurers need a fast, reliable way to separate what the storm actually caused from what was already there. The clearest check is historical aerial imagery, timestamped before the loss date, showing exactly what the roof looked like before. That check only works if the imagery archive goes back far enough. Without deep historical coverage, pre-existing damage is invisible, and carriers end up paying for fraudulent claims.
)
Stale property intelligence stops protecting the decisions it was built for. Every guarantee behind the data has to be actively maintained against a world that keeps changing.
What’s already in the works
Most maintenance happens inside the platform, powering silent improvements. However, there are times when the curtain gets pulled back. That’s because we want your reaction before it ships, not after.
That’s Nearmap Labs, a new public innovation hub giving customers, partners, developers, and industry leaders an early look at the technologies and research Nearmap is exploring to advance property intelligence. Some of it is early exploration. Some of it is close enough to test right now. None of it is finished yet, and that’s the point.
Here’s what Nearmap Labs looks like today:
Zero-Shot Search finds specific property conditions in imagery without training a dedicated model for each one. Normally, if you wanted an AI system to reliably spot solar panel installations across a portfolio, someone had to build and train a model for that exact task. Zero-shot search lets a model recognize conditions it was never explicitly trained on, which matters most when a new risk category shows up faster than a traditional model could be built to catch.
Temporal Reasoning Engine is built around the question of whether what you’re seeing is real change or just noise. A roof looks different in June than it did in January for a dozen reasons that have nothing to do with damage. This tool reads a full imagery timeline instead of comparing two snapshots, so it can tell a genuine structural change apart from seasonal variation, shadow angles, or capture conditions. For anyone assessing property condition over time, that’s the difference between a real signal and a false alarm.
The National Roof Age Study looks at how roof condition connects to the climate pressure a property truly faces, market by market. One roof score tells you about one property. Laid across the country, that same data starts to say something about where climate risk is concentrating, where replacement cycles are accelerating, and where claims activity is likely headed next.
Insurance Rules Mapper takes the mapping rules and thresholds carriers already use and connects them directly to Nearmap AI attributes. Right now, a lot of that mapping happens by hand, carrier by carrier, rule by rule. Automating it means your underwriting rules are faster to configure against Nearmap attributes, and easier to keep consistent.
And two integrations, Nearmap for Autodesk and Nearmap for ArcGIS, put Nearmap imagery and analytics directly inside Civil 3D and ArcGIS. Now, AECO and GIS teams don’t have to export, re-import, or leave the software they already live in to get to it.
Early access to these tools is open
The tools are all works in progress. Some might change a lot before they ship. Others may not ship at all in their current form. That’s normal for anything still being tested against real use cases — and it’s why these are all visible now instead of only after they’re polished.
You’re not just watching this happen
Most companies only show you all the cool stuff they’ve already built. That framing puts you on the outside, watching a company innovate as you passively wait for the next update.
That’s not how Nearmap works. The reason Insurance Rules Mapper exists in its current form is that underwriters dealt with the same manual mapping problem. The reason Temporal Reasoning Engine focuses on separating real change from noise is that property claims teams kept running into false positives from seasonal variation.
The tools in early access today aren’t our idea of what customers might want. They’re built from what customers have already said they need, refined through direct testing before general release.
That means the way you use property intelligence today — the questions you ask of it, the gaps you notice, the workarounds you build because something isn’t quite right yet — drives what gets built next.
Think about how the National Roof Age Study came together. Roof age was already an attribute Nearmap tracked property by property, for individual underwriting and inspection use cases. It became a market-level study because carriers still had questions that needed answers. More than wanting to know the age of individual roofs, carriers looked to understand how roof age varies across an entire region, and what that can predict about upcoming claims volume. We had the data. But the research direction came from people asking a bigger question than the tool was originally built to answer.
)
Coming down the pipeline
This is how new capabilities tend to move here. Research turns into a tested tool, a tested tool earns its way into early access, and early access becomes a standard part of the platform once it holds up against real conditions.
That pattern isn’t slowing down. New tools are surfacing every quarter, several built as direct answers to problems customers have already put into words. Imagery keeps getting sharper. AI keeps closing the gap on conditions it currently misses. Materials pricing keeps tightening against cost shifts instead of trailing behind them.
You have a say in which of these becomes something you rely on. Use the platform the way you already do. Speak up the moment something doesn’t quite work. That kind of feedback has a track record of becoming the exact tool you asked for, faster than you’d expect.