The AI Model Builder has no access to data from SimaPro and, thus, never sees your models, libraries, account details or any other data in SimaPro. Only your four wizard answers (your product description, the two geographies, the life cycle area and the reporting standard) are sent as input. Every read and write against the platform is performed by our own backend service, not by the AI.

No. Your inputs are not used to train or improve any SimaPro model or algorithm. That is kept entirely separate.

Only the SimaPro team can turn this feature off at account/organization level. Please contact [email protected] to request any changes. When it is off, no AI element appears anywhere in the interface for that account.

They stay, and they remain usable exactly as before. Once a model is built it is an ordinary SimaPro model, so turning AI Model Builder or any other AI feature off only removes the way you create new ones. Nothing already in your organization is changed, hidden or restricted by enabling or disabling the feature.

There is no cap on the number of models and no credit system.

There are two practical constraints. You generate one model per wizard session, and the wizard only appears on an empty canvas, so you cannot generate into a project that already has processes in it.

The AI determines the process hierarchy, the process names, the reference amounts and the quantities on each flow, guided by system prompts that encode generic modelling rules and the reporting standard you selected. Those modelling rules were written and verified by SimaPro’s own LCA consultants and in-house experts, so what the AI follows is PRé Sustainability’s modelling practice rather than a general-purpose model’s own idea of how an LCA should be structured.

The system boundary comes from your Step 3 answer. You choose either Production or Production, Use and End of life, and the AI generates the life cycle stages that scope implies. Each generated process carries a reference amount and unit, shown on the confirmation screen before you build. The reference amounts are built relative to the product you input in Step 1.

Yes. The output is a real, editable SimaPro project, not a preview or a mock-up, and every standard action is available on it. The AI Model Builder generates foreground processes only. Linking them to background databases is a manual step you perform after the build. A calculation run before those links are made will not return meaningful results.

You review before you commit, not after. The confirmation screen shows the complete draft as a collapsible hierarchy, with the name, reference amount, unit and geography on every process, alongside the four answers you gave. If the structure is wrong, you go back through the wizard and regenerate rather than fixing it by hand.

After the build, the post-build summary lists what was created and flags what still needs manual completion. What you are checking is the structure, not the numbers.

You cannot tell from looking at the model. There is no persistent visual indicator on an AI-generated model; it is on the list for a future release rather than in the current one.

Scope is at least unambiguous: because the wizard only runs on an empty canvas, everything in the initial build came from the generator, and everything after that is your own editing. There is no mixed state within the generated portion.

Screening, learning and early hotspot identification: it is a structural starting point for exploration. A published study needs the background data linked, the inventory filled with real data, and expert review, none of which this feature does.

You do. A model generated with AI Model Builder belongs to the user or organization that created it, on exactly the same terms as a model you built by hand. SimaPro claims no rights over it and places no restriction on commercial or client use.