Ask a planner, an infrastructure strategist or a policy adviser what slows them down, and you rarely hear "we don't have the data". You hear "I can't get to the answer". The data is there. The GIS is there. What's missing sits between them.
This article sets out how we think about that gap, why we call it the Spatial Evidence Layer, and why we believe it's the layer Geographic information systems (GIS) were never inherently designed to provide.
The hard part happens before the map
Every place-based decision rests on a deceptively simple set of question such as: what's true about this place? Which sites are suitable? Who lives near the proposed services? Who or what's exposed to a hazard? Where does capacity exist and where it doesn't?
Answering well means assembling evidence. In practice that looks like this:
Find the relevant datasets, across open data portals, state and federal services, internal systems and spreadsheets.
Understand each one: who published it, when, at what resolution, with what gaps.
Reconcile boundaries, projections, formats and definitions so the datasets can be combined.
Analyse them: buffers, joins, intersections, cross-tabs etc….
Explain the result to someone who has to approve, fund or defend the decision.
Repeat it all when the data changes, or when a colleague asks the same question about a different place.
Most of that work isn't analysis. It's finding, checking and reconciling. And most of it is repeated, by different people, in slightly different ways, across the same organisation. These are commonly referred to as workflows.
Why GIS alone can't close the gap
None of this is a criticism of GIS. Geographic information systems are among the most capable tools in any organisation. They manage, edit, analyse and map spatial data with a rigour no other technology matches, and they're the system of choice for a reason.
But GIS systems were built to manage and map data, not to make spatial evidence quick to assemble or easy to verify. They assume a trained operator, a prepared dataset and a defined task. They were not designed for a transport planner who needs an answer by Thursday, or a policy adviser who wants to test three scenarios before a briefing due in 30 minutes.
So organisations end up in one of two places. Either every spatial question queues for a small GIS team, which becomes a bottleneck through no fault of its own. Or domain professionals work around the queue, pulling together evidence by hand in spreadsheets and slide decks, where nobody can fully trace how a number was produced.
The first path is slow. The second is risky. Both leave most of the organisation's spatial evidence locked away from the people who need it, and both have significant productivity implications.
The place-based intelligence stack
It helps to see the whole stack. Place-based intelligence in a modern organisation has five layers:
Capture. Surveys, sensors, imagery, administrative records.
Data and infrastructure. Open data portals, databases, cloud storage, data platforms.
GIS systems of record. ArcGIS, QGIS, spatial databases: where data is managed, edited and mapped.
Spatial Evidence Layer. Where evidence is discovered, assembled, validated and explained.
Decisions and workflows. Plans, business cases, briefings, approvals.
Layers 1 to 3 are mature. Layer 5 is where the organisation lives. Layer 4 is where most of the friction sits, and until recently it has been done by hand: by GIS analysts fielding requests, and by domain professionals stitching evidence together themselves.
That's the layer MapAI is built for.
What a Spatial Evidence Layer does
A Spatial Evidence Layer has four jobs.
Discover. Find the spatial evidence relevant to a question, across open data, licensed data and the organisation's own holdings, and make clear what each source is and isn't.
Assemble. Bring those sources together: align boundaries and formats, combine them, and run the spatial analysis the question needs.
Validate. Make the method inspectable and repeatable, so the same question over the same area gives the same answer, and so a GIS expert can review and approve how it was done.
Explain. Deliver the result with its working attached: sources, assumptions, steps and gaps. An answer nobody can explain isn't evidence.
Then it delivers two things to two audiences. Domain professionals get trusted answers. GIS experts get analysis-ready data and methods they control.
What it must not become
Defining a category also means saying what it isn't.
A Spatial Evidence Layer is not a GIS replacement. It sits alongside GIS and depends on it. The system of record stays where it is.
It's not a replacement for expertise. Interpretation, uncertainty and accountability stay with people. The layer automates the mechanical work so expertise is spent where it matters, in areas of innovation, insights and impact.
And it's not "chat with a map". Natural-language interfaces are quickly becoming a feature of every spatial tool. A conversational front end on top of an opaque process just moves the problem. The value is in what sits behind the conversation: the data intelligence, the repeatable methods and the evidence trail.
Why "trusted" has to be earned
We describe MapAI's position as trusted spatial answers. We're careful about that word.
Trust isn't a claim a vendor gets to make about itself. It's earned through reliable data, transparent methods, validation and provenance, question by question. The job of a Spatial Evidence Layer is to make trust possible: to show enough of the working that a professional can decide for themselves whether to stand behind the answer.
That's the standard we hold ourselves to. It's also, we think, the standard any organisation should hold any AI-assisted spatial tool to.
Where this goes
Over the next few years, more and more spatial questions will be asked by people who aren't GIS specialists, using natural language, often inside tools that don't look like GIS at all. That's a good thing, if the evidence behind the answers holds up.
The organisations that get this right won't be the ones that replace their GIS teams with chatbots. They'll be the ones that give their GIS experts a way to validate methods once and let the whole organisation reuse them, and give their domain professionals trusted spatial answers they can inspect, question and defend.
That's the gap between GIS and decisions. It's what the Spatial Evidence Layer is for.
Try MapAI on your own question
Sign up and start free, or talk to us about a structured trial for your team.
Sign upChris Pettit is a Co-founder of MapAI, and Scientia Professor and Director of the City Futures Research Centre UNSW.