Spatial intelligence guide
From Prompt to CAD: How to Export Sourced AI Site Analysis Back to Rhino & AutoCAD
AI has made site analysis dramatically faster. But for architects, speed is only useful if the result can become part of the design workflow.
A beautiful map or AI-generated image is not enough. Architects need editable geometry: street networks, building footprints, site boundaries, green areas, transport infrastructure and other spatial layers that can move from analysis into the tools where design actually happens.
That is where the difference between AI-generated imagery and AI-powered spatial analysis becomes important.
Aino lets architects describe the analysis they need in plain language, work with sourced spatial data, and export the resulting geometry as CAD-ready files. Building footprints, street networks and site boundaries can be exported as geo-referenced DXF or GeoJSON and opened in tools such as Rhino and AutoCAD.
The result is a different workflow:
Prompt → spatial data → analysis → editable geometry → CAD
Instead of:
Prompt → PNG → redraw everything manually
Why AI site analysis often breaks the architectural workflow
The first generation of AI tools for architecture focused heavily on images.
You could generate:
• a site concept;
• a streetscape;
• a massing study;
• a contextual rendering;
• an aerial visualization.
These outputs can be useful for exploration, but they have a fundamental limitation:
pixels are not geometry.
A PNG does not give you an editable street centerline.
A rendered building does not give you a building footprint.
An AI-generated aerial image does not give you reliable parcel boundaries.
And a beautiful site-context image cannot be snapped, measured, edited or used as a base layer for a CAD model.
For architects, the useful question is therefore not:
Can AI generate a map?
It is:
Can AI generate spatial information that I can actually continue working with?
That distinction is at the heart of an AI site analysis workflow.
From natural language to spatial layers
Traditional site analysis often starts with a long list of manual tasks.
An architect or GIS specialist may need to:
Find the project location.
Download base mapping.
Find building footprints.
Find roads and paths.
Collect parks and green areas.
Identify relevant transport infrastructure.
Clean and combine datasets.
Reproject the data.
Import it into GIS or CAD.
Style the layers.
Clip the analysis to the project area.
Export the result.
The analysis itself may take minutes.
Preparing the data can take hours.
An AI spatial workflow changes the starting point.
Instead of translating the design question into a series of GIS operations, the architect can describe the spatial context they need.
For example:
“Show the building footprints, street network, parks and green areas within 500 metres of the site. Keep the layers separate and prepare the result for CAD export.”
The important part is that the request is not asking an AI model to draw what it thinks the city looks like.
It is asking the system to work with spatial data.
That makes the output fundamentally different.
AI-generated image vs. AI-generated spatial data
Consider two approaches to the same project.
Approach 1: Image generation
You prompt an image model:
“Generate an aerial architectural site analysis showing buildings, roads, parks and surrounding context.”
You get an image.
It may look convincing.
But if the road is in the wrong position, there is no underlying geometry to correct. If a building footprint is inaccurate, you cannot simply edit the polygon. If you want to use the road network in Rhino, you have to recreate it.
Approach 2: Spatial analysis
You ask an AI spatial system:
“Show building footprints and the street network around this site.”
The system retrieves and processes spatial datasets.
The result consists of geographic features.
Those features can then be:
• inspected;
• filtered;
• styled;
• analysed;
• exported;
• reused in another project.
This is the critical distinction.
Generative imagery approximates context. Spatial analysis works with context.
A practical architect’s workflow
Let’s take a typical early-stage architecture project.
You have a site address and need enough context to begin massing studies.
The first question is not yet:
“What should the building look like?”
It is:
“What is actually around this site?”
You need the physical context first.
Step 1 — Start with the site
Begin with an address or project location.
Aino can turn the location into the starting point for a spatial project, rather than requiring you to assemble a GIS workspace manually.
From there, define the scale of the analysis.
For example:
• 100 m for immediate site context;
• 300 m for the surrounding urban fabric;
• 500 m for local amenities;
• 1 km or more for transport and broader urban context.
The correct radius depends on the question you are trying to answer.
Step 2 — Ask for the layers you actually need
Instead of downloading everything, ask for the relevant spatial context.
For example:
“Show building footprints, roads, pedestrian paths, parks and green spaces within 500 metres of the site.”
For an architecture project, you might then add:
“Show the main public transport stops and stations within 800 metres.”
Or:
“Identify buildings and open spaces immediately surrounding the site.”
The advantage is not simply that the map appears faster.
The advantage is that the spatial analysis starts from the design question rather than from a catalogue of GIS operations.
Step 3 — Keep the layers separate
This matters once the analysis moves into a design environment.
A site-context map should not become one flattened graphic.
You may want:
• roads on one layer;
• building footprints on another;
• parks and green areas on another;
• site boundary separately;
• transport infrastructure separately.
That gives the architect control over what gets imported, displayed, edited or discarded.
It also makes the exported information more useful downstream.
For example, in Rhino you might want to:
• hide all buildings;
• keep only surrounding streets;
• isolate green space;
• use building footprints as reference geometry;
• create your own design layers on top.
The spatial analysis becomes a starting point for design, not the final presentation.
Step 4 — Check the sources before you design
Speed should not mean giving up traceability.
A professional site analysis needs to distinguish between:
analysis generated from sourced spatial data
and
AI-generated assumptions.
This is particularly important when the output will influence:
• a planning proposal;
• a feasibility study;
• an investment decision;
• a site acquisition;
• a design submission.
Aino’s positioning is built around sourced and structured spatial analysis rather than treating an AI model’s visual output as ground truth. Its site-analysis workflow connects data gathering, analysis and visualization, while the platform highlights source-tracked findings as part of the workflow.
That distinction should remain clear throughout an architectural workflow:
AI can automate the analysis. The architect remains responsible for interpreting the result.
Step 5 — Export the geometry
Once the analysis is ready, the next question is simple:
Can I take this data with me?
For architectural workflows, Aino supports export of spatial outputs including DXF and GeoJSON. Its CAD export workflow is specifically described for building footprints, street networks and site boundaries, with geo-referenced output ready for Rhino or AutoCAD.
This changes what “AI site analysis” means.
The output is no longer just:
a report
or
a screenshot
but potentially:
editable spatial geometry.
DXF vs. image export
The difference is worth making explicit.
· Output · Useful for · Editable geometry ·
· --- · --- · --- ·
· PNG · presentations, references · No ·
· PDF · documentation, presentations · Limited ·
· SVG · vector graphics and diagrams · Yes, depending on workflow ·
· GeoJSON · GIS and web mapping · Yes ·
· DXF · CAD workflows · Yes ·
If the next step is architectural design, DXF is often much more useful than a raster image.
A DXF can become part of the working drawing environment instead of sitting beside it as a reference image.
Bringing the site context into Rhino
A typical workflow might look like this:
Analyse the site in Aino
Start with the project location and generate the spatial layers you need.
Export the relevant geometry
Export the site context as DXF.
Open the file in Rhino
Import the DXF into your Rhino project.
Check units and coordinates
Before using the geometry for design, confirm the coordinate system, units and project origin are appropriate for your workflow.
Organize the layers
Separate roads, buildings, green spaces, site boundaries and other contextual elements.
Begin design
Now the context can be used directly for:
• massing;
• setbacks;
• circulation studies;
• access studies;
• public-realm design;
• site organization;
• early urban design.
The important change is that the AI analysis does not end when the map is generated.
It continues into the design environment.
Bringing the same analysis into AutoCAD
The principle is similar.
Export the relevant site geometry as DXF and bring it into the project drawing.
From there, the architect can continue with the existing CAD workflow:
• layer management;
• dimensions;
• annotations;
• blocks;
• linework;
• design overlays;
• site planning.
This makes AI spatial analysis compatible with a workflow that architects already know.
The objective is not to replace CAD.
It is to remove the repetitive preparation that happens before CAD becomes useful.
The real productivity gain happens before design
Consider a simplified project timeline.
Traditional workflow
Site address
↓
Search for data
↓
Download GIS layers
↓
Clean data
↓
Reproject
↓
Clip layers
↓
Style layers
↓
Export
↓
Import into CAD
↓
Start design
AI-assisted spatial workflow
Site address
↓
Prompt
↓
Sourced spatial analysis
↓
Review
↓
DXF / GeoJSON export
↓
CAD
↓
Design
The architectural thinking remains.
The repetitive spatial preparation is reduced.
That distinction is important.
AI should not replace the architect’s judgment.
It should reduce the amount of time the architect spends doing work that does not require architectural judgment.
What you can export for an early site study
A useful starting context might include:
Physical context
• building footprints;
• roads;
• paths;
• site boundaries;
• open spaces;
• parks;
• green areas.
Urban context
• public transport;
• amenities;
• commercial uses;
• surrounding development;
• key destinations.
Analytical layers
• accessibility zones;
• buffers;
• isochrones;
• density;
• demographic context;
• other spatial indicators relevant to the project.
Aino’s broader site-analysis workflow is designed around combining these types of layers and turning the findings into presentation-ready maps or exported spatial data.
Prompt examples for architects
The easiest way to start using AI spatial analysis is to formulate prompts around a design question.
Basic site context
“Show all building footprints and roads within 500 metres of this site.”
Urban context
“Show buildings, streets, parks, green spaces and public transport around the site.”
Immediate surroundings
“Map the buildings and open spaces within 200 metres and keep each layer separate.”
Accessibility
“Create a 10-minute walking isochrone from the site and show the main amenities inside it.”
Design constraints
“Identify the major roads, public spaces, green areas and surrounding building footprints that could influence the site’s urban context.”
The important principle is:
Ask for spatial information, not an image of spatial information.
Why geo-referencing matters
A map can look correct and still be practically useless if its geometry does not align with the rest of the project.
Geo-referenced export means the spatial features retain their geographic position so they can be placed correctly in a broader spatial workflow.
For architects, this is especially important when combining:
• site survey information;
• cadastral data;
• GIS layers;
• aerial imagery;
• CAD drawings;
• infrastructure data.
The goal is not simply to export a shape.
The goal is to export a shape that still knows where it belongs.
AI does not eliminate the need for verification
There is an important professional caveat.
An AI-assisted site analysis should not automatically be treated as a legal or survey-grade source.
Building footprints can be outdated.
Road data can change.
Open datasets can have different update cycles.
Planning regulations may require authoritative municipal sources.
Survey data may be more precise than publicly available spatial datasets.
Therefore, the right workflow is:
AI for research and analysis
→ source verification
→ professional interpretation
→ design decision
For high-stakes decisions, always verify critical geometry and regulatory information against the authoritative source.
This is also why source-tracked spatial analysis matters: the architect needs to know not just what the system returned, but where the underlying information came from.
From site analysis to a reusable design workflow
The biggest opportunity is not exporting one DXF.
It is building a repeatable workflow.
Suppose an architecture studio frequently works on urban infill projects.
The team could standardize a site-context analysis containing:
• building footprints;
• roads;
• public transport;
• parks;
• amenities;
• demographic context;
• accessibility.
Instead of recreating the same GIS workflow for every project, the studio can reuse the analysis structure and layers.
Aino is designed around this idea of reusable layers, analysis templates and shared project workspaces.
That turns site analysis from a one-off task into part of the firm’s project infrastructure.
The bigger shift: from AI images to AI geometry
The architectural AI conversation has been dominated by image generation.
But the more useful transition may be happening somewhere else.
AI-generated images help architects explore possibilities.
AI-generated spatial analysis helps architects understand constraints and context.
AI-generated geometry can connect that analysis directly to the tools where design happens.
That creates a more useful chain:
Question → Data → Analysis → Geometry → Design
rather than:
Question → Image → Manual redraw
For professional architecture workflows, that distinction is significant.
Final takeaway
The most useful AI site analysis tool is not necessarily the one that produces the most impressive image.
It is the one that can answer a spatial question, show where the answer came from, and give you something you can actually use afterward.
For architects, that means moving from:
“AI made me a map.”
to:
“AI prepared the spatial context for my project.”
With Aino, the workflow can start with a natural-language prompt, continue through sourced spatial analysis, and end with geo-referenced geometry that can be exported into the CAD environment — including DXF workflows for Rhino and AutoCAD.
The goal is not to replace Rhino or AutoCAD.
It is to make sure you spend more time designing in them — and less time preparing the map before you can start.
Try the workflow
Start with an address, describe the spatial context you need, review the resulting layers, and export the geometry you want to continue working with.
From prompt to spatial analysis. From spatial analysis to CAD.
FAQ
Can AI spatial analytics be trusted?
AI spatial analytics can be trusted when results are tied to real data sources, clear spatial operations and evidence that a professional can verify.
Why does source tracking matter for spatial analysis?
Source tracking shows where a result came from, which dataset was used and how the finding was produced, reducing the risk of black-box AI outputs.
How does Aino support verifiable spatial analysis?
Aino uses AI to orchestrate spatial workflows while keeping the underlying data, findings and provenance visible for review.
What practitioners say
“Exporting sourced layers back into CAD saved hours and kept the analysis editable.”
— Matteo, Computational Designer
Further reading and resources
Continue with related Aino guides, product pages and documentation: