GenAnywhere Generate Now

Select a map area. Generate a 3D city model with AI.

No drone survey required. Turn map data into editable 3D buildings, terrain, water and vegetation for BIM, GIS and 3D workflows.

Generate now

Site model limits: what it can and cannot show, GenAnywhere

The useful parts of a generated scene come with edges, and the edges are worth naming before a review.

Site model limits follow from the source and from the way a scene is generated. GenAnywhere builds buildings, terrain, water and vegetation from map and satellite data, with estimated heights and inferred types. That makes a strong instrument for arrangement and a poor one for measurement, and this article separates the two.

Aerial view of a low-rise neighbourhood with gardens and streets

1 Where limits begin

1.1 What does the source decide?

The source decides what exists in the scene and what is missing from it. A road built after the satellite pass will not be there, and a building demolished last year may still stand. No amount of editing removes that starting point, and the date is the reader's way to judge it.

The source also decides how much the scene can say about the edge of a site. Coverage thins at the boundary of a dataset, so a neighbouring block can arrive half drawn. Extending the area is the fix, and checking the edge is the habit.

A source can also be right and still mislead. A site that has not changed for a decade may be well represented by older data, while a district under active construction can be wrong within months. The date carries more meaning than the age alone.

1.2 What does the tool estimate?

The tool estimates height, infers building type and generates a vegetation band. Those values are plausible at district scale and unreliable at the scale of a single property. A warehouse and a tall shed can share the same footprint and end up with similar heights.

Estimation is not a defect. It is the trade that makes a scene appear without a survey. The defect is forgetting the trade at the moment a number is quoted in a meeting.

The same applies to water and vegetation. A river edge follows the map, and a tree belt follows the pattern in the source. Neither is a flood line, a species record or a reason to skip a site visit.

2 What it shows well

2.1 What about massing and scale?

Massing is the strongest use. The scene shows how a proposal sits against its neighbours, how the skyline changes and whether a block is too deep for the street. Those judgements do not need measured heights, because they are about proportion.

A team can test two options in an afternoon and see which one holds the street edge. That is a design conversation, and the scene is built for it.

The scene is also good at showing what a proposal hides. Raised ground, a new podium or a taller neighbour can change a view that looked settled in plan, and the change appears as soon as the model is turned.

2.2 What about streets and enclosure?

Streets, frontages and enclosure read well in a scene with buildings and ground. A narrow lane feels narrow, and a wide junction feels wide, even when every dimension is approximate. The eye picks up the relationship rather than the number.

The scene also shows access in a practical way. A service yard that looks tight on a plan can look impossible in three dimensions, and the reverse happens too.

Walk the route in the viewer before the meeting. A turning circle that works on paper can fail once a building has depth, and that failure is cheaper to find in the scene than on site.

Team reviewing a site model around a meeting table

3 What it cannot show

3.1 What about condition and structure?

Condition, structure, services and floor levels are outside the scene. A building that looks sound in the model can be derelict, and a plot that looks empty can hold foundations. The scene describes shape and says nothing about what the shape is made of.

Professionals who assess condition work to published guidance, such as the RICS professional standards, and a generated scene does not stand in for that work. It can point at the buildings that need a closer look.

3.2 What about legal and measured facts?

Boundaries, ownership, rights of way and measured levels are outside the scene. So are planning constraints and any question that depends on a legal record. The model can frame those questions and cannot answer them.

This is the limit that causes the most trouble, because a boundary drawn from a map looks authoritative. Treat every line as a representation until a legal or measured source confirms it.

The same caution applies to rights of way and to any easement. A path that appears in the source may be a track rather than a public route, and the scene cannot tell the difference.

4 How limits surface in a review

4.1 What does a confident wrong answer look like?

A confident wrong answer is a number quoted from the scene without its source. A height read off the model sounds like a measurement, and it travels into a report where nobody can trace it. The scene did not lie; the sentence did.

The same happens with area. An outline drawn from a map is not a surveyed boundary, and a figure calculated from it inherits that weakness. If the figure matters, it needs a measured source.

Confidence also spreads through images. A rendered still looks like a proposal, and a viewer may read it as a forecast rather than an arrangement. Label the image with the stage it belongs to, and the reading stays closer to the truth.

Large site model on a table with rolled drawings

4.2 Which question exposes it?

Ask what the number would change if it were ten percent different. If the decision survives, the estimate is good enough for the conversation. If the decision changes, the scene has reached its limit and another source is needed.

That question is quick and it works in a meeting. It turns a debate about numbers into a debate about consequences, which is usually the more useful one. It also gives the team a way to stop arguing without dismissing the concern.

5 How to state the limits

5.1 What does a scope note contain?

A scope note says what the scene is for, what it was built from, when the source dates and what it does not cover. Four lines are enough. The note travels with the file so the reader does not have to ask.

Keep the note next to the scene rather than buried in a folder. A note nobody opens is the same as no note at all. Name the file so the purpose survives a change of staff, and keep the date visible in the name where the tool allows it.

5.2 Which sentence should a team use?

A workable sentence is: this is a context scene generated from map data for arrangement studies, not a measured record. It is plain, it is short, and it sets the expectation before questions arrive.

Adapt the wording to the audience, and keep the meaning. A client, a planner and a colleague all need the same boundary, expressed in their own language.

6 Living with limits

6.1 When does the limit matter?

The limit matters when a decision depends on a measured fact. A level, a boundary, a structural span or a legal constraint will not be settled by the scene. In those moments, the model should step back and the survey should step forward.

Start with a small area in the GenAnywhere app when the question is about arrangement, and stop before the question turns into measurement.

6.2 When does it not matter?

The limit rarely matters in a massing study, a consultation image or an early options review. Those conversations are about relationships, and relationships survive approximation. A team that names the limit can use the scene with confidence rather than apology.

6.3 When should another tool take over?

Another tool takes over when the work needs authored elements, documented properties or measured geometry. A design or BIM tool carries that structure, and a survey carries the measurement. The context scene stays useful as a reference beside them.

Review the scene limits in the GenAnywhere workflow.

7 FAQ

FAQ-1 Is the scene accurate enough to measure?

No. Heights are estimated and outlines follow a map source, so a dimension taken from the scene carries both weaknesses. It is fine for proportion and for conversation, and it is not a measured record. If a decision depends on a dimension, commission a survey and use the scene to explain the brief.

FAQ-2 Can I rely on the boundary?

As a representation, not a record. The boundary in the scene follows the source geometry, and it is not a legal or surveyed line. Check any boundary that matters against the title or a survey, and keep the scene out of that particular argument. If the boundary is the subject, the scene is the wrong instrument.

FAQ-3 Does it show recent changes?

It shows what the source captured. A road, a building or a cleared plot that appeared after the source date will be missing or out of date. Check the date, make the edits that matter, and record them so the next reader knows what changed. A scene without that note invites assumptions the source cannot support.

FAQ-4 When should I avoid it?

Avoid it when the question is about measurement, condition, structure, ownership or compliance. Those questions need a qualified source, and a generated scene can point at them. Use the scene for arrangement and context, and bring in the right instrument for the rest. The scene remains useful as a way to frame the request.

Content and technical notes

This article was prepared by the GenAnywhere team with AI-assisted drafting and human review.

Product features, formats, and availability change as the platform evolves; refer to the official GenAnywhere website for current details.

For technical questions or corrections, contact support@genanywhere.com.

Last reviewed: 2026-11-06.