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

Model alternatives: a comparison checklist GenAnywhere supports

A short list for choosing between site-model options without turning the choice into a procurement exercise.

Model alternatives are easier to compare when the checklist follows the evidence rather than the feature list. The questions below cover the source, the edit, the export and the handover, which are the points where one option quietly costs more than another. GenAnywhere supports the checklist by supplying an editable base and a general GLB, but the questions apply to any method.

Aerial view of a hillside town with a road winding below

1 Set the comparison

1.1 What is the comparison for?

The comparison exists to choose an instrument for a question, not to rank tools. Start with the decision the scene must support.

Write the decision in one sentence. Every option that cannot support it drops away, and the remaining comparison is shorter. A question with two parts usually produces a scene that serves neither, so split it before choosing a method.

Keep the sentence at the top of the pack, so the review has something to return to when the discussion drifts.

Then set the area. A comparison run on a site that is too large will reward whichever option produces the most geometry rather than the one that answers the question. Draw the boundary at the edge of the decision, and add neighbours where they change an argument.

1.2 Who sets the criteria?

The user, the payer and the maintainer. Each cares about a different property, and a criterion owned by nobody tends to be ignored.

Agree the criteria before scoring. A checklist that changes after the answers is a way of reaching a decision that was already made.

Write the criteria down with the question. The pair is what makes the review repeatable. Criteria that are not written down are usually applied inconsistently.

2 Check the source

2.1 Where did it come from?

Map data, a flight, a survey or drawings. The checklist should name the source and its date, because a scene without either cannot be tested. A source with a date can be checked against the site, and the date is the cheapest check in the list.

Check that the source covers the site and the neighbours. A base that stops at the boundary leaves the context out, and the context is often the argument. A coverage drawing belongs in the package for the same reason.

Metadata conventions such as those published by the FGDC are useful when a record has to outlive the team that made it.

Physical planning model of a district beside a river

2.2 What does it not cover?

Condition, measured levels and recent change. Each of those needs another record, and the checklist should say so. A gap with an owner is easier to close than one that is merely acknowledged.

Name the gap rather than filling it with a guess. The checklist is there to make gaps visible, not to hide them.

A scene that admits two gaps is easier to trust than one that appears complete.

3 Check the edit

3.1 Can it be edited?

Try one edit before scoring: move a block, change a height, replace a building. The response tells you how the scene behaves under revision. An edit that takes a minute is a different instrument from one that takes an hour, and a slow scene is used less often.

A base that keeps its layers is editable; a mesh that carries no separate parts is not.

The test takes a minute and says more than a feature list.

3.2 Is the edit recorded?

Every replacement or height change should leave a note. An unrecorded edit makes the scene harder to explain later. A single line per change is enough for a later reader to follow it.

The record also protects the original source. A reader can tell what came from the map and what the team changed.

A short edit log is enough. It does not need to be a document.

4 Check the export

4.1 What does it carry?

Geometry and materials, in a general container. Layers and names may or may not survive the trip.

Test the export in the receiving tool. A file that opens and measures correctly is worth more than one that merely opens. A file that opens but measures wrongly is more dangerous than one that fails.

Check one known dimension after import, and check the layer names at the same time.

Aerial view of a bay city between hills and water

4.2 What is lost?

Authored meaning. The receiver gets shapes rather than elements with properties, and the difference matters for coordination.

Say which loss matters for the use. Context tolerates it; coordination does not. Expectations set before the handover prevent a disagreement after it.

A sentence in the note is enough to set the expectation, provided the sentence names which loss is acceptable and which one is not.

Ask the receiving team to read the note before opening the file. A reader who expects shapes will not spend an afternoon looking for properties that were never exported.

5 Check the handover

5.1 What travels with the scene?

Source, date, coverage, edits and checker. Five items on one page, and the receiver can work without calling.

A package that opens without a call is the goal. The note is what makes that possible.

The package should name a contact for the questions that remain. A contact is cheaper than a search through old messages, and it gives the receiver a route that does not depend on one person's memory.

5.2 What does the receiver confirm?

One dimension, the layers, the coverage and the note. Four checks, a few minutes, and a record of the result.

Record the result so the next team starts from a known baseline.

If a check fails, stop and ask rather than defending the file. The failure points at the part of the scene that needs another source, and it is information rather than embarrassment.

6 Close the review

6.1 What makes the decision defensible?

A stated question, a recorded source and a note that says what was and was not covered. Those three items explain a choice to anyone who joins later.

The record of rejected options also helps. It explains why the chosen one was chosen, which is often the harder question. One line per rejected option is enough to show the reasoning.

A short entry is enough. The decision does not need a report.

Defensibility is mostly about consistency. If two reviews of similar sites reach different conclusions, the criteria should explain the difference rather than the personalities in the room.

6.2 When is the checklist too long?

When it takes longer than the decision it supports. A short list that is run beats a long one that is skipped.

Cut any check that has not changed an answer in several projects. Length is the main reason a checklist stops being used, and a list that grows after every project becomes a document nobody reads.

Keep the checks that catch the expensive mistakes, and drop the rest.

6.3 When does the review close?

When the scene answers the question and the note records the limits. The closing step is a short entry, not a report.

Close with the records kept and the next question named. That is enough for the next team to pick up.

A closed review is easier to reopen than an open one that never ended. The closing note should name the next question, so the decision becomes a handover.

Draw the area in the GenAnywhere app and run the same checklist before the scene reaches a reviewer.

7 FAQ

FAQ-1 Is this checklist about software?

No. It is about evidence and cost. The software follows from the answers, not the other way round. Two teams using the same tool can still need different methods, and the checklist is designed to show why. The evidence should drive the choice, and the tool should follow it. A checklist that begins with products tends to end with a purchase instead of a decision.

FAQ-2 How many options should it cover?

Three or four. Include the method the team already uses, because it is the baseline every alternative is judged against. More options add cost to the review without necessarily improving the decision. A longer list tends to favour the option with the most features rather than the right evidence. Drop any option the team would not actually use.

FAQ-3 Does it work for a single site?

Yes. The checklist scales down as well as up. A single site may need just the source, the edit and the handover checks, and the rest can be left out without weakening the decision. Those three checks carry most of the value. Add the export check when another team receives the file, and leave the rest for larger sites where the cost of a wrong choice is higher.

FAQ-4 What is the most common failure?

The handover. Teams spend effort on the scene and little on the note, so the receiver cannot tell what was checked. A one-page note with the source, date, coverage and checker prevents most of the trouble. A short note is cheaper than a rebuild and easier to defend. Write it while the scene is still open, not after the questions arrive.

See how a map area becomes an editable site model in the GenAnywhere workflow.

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.