We never had software that would handle the cradle to grave job - that is a full integrated stack including suppliers like Detroit's Tier 1, Tier 2, where we could dictate what software the suppliers and internal users used.
We also never had a customer who could take delivery solely of such models.
Even so I built all my models so all the geometric data and feature controls were part of the model and were "shown" on the drawing rather than being created in the drawing mode. This was key for using the VSA tolerance analysis software as it worked directly from the model information.
There were large gaps in the software - I don't recall any means of recording finish requirements in a model so that one could select a surface and have it tell the class of conversion coat, the primer, the allowable range of primer thickness, or the top coat and range of thickness and so forth or select the part marking and it reports the epoxy ink color and spec. Recording those in the model were on the big list I tried to get ComputerVision to implement, especially the vexatious paint masking requirements.
Lucky me and primarily government contracts meant the required deliverables to them was mostly PDFs; the USPS also wanted the native files, the PDFs, and DWG files because they used AutoCAD for everything. Suppliers most always wanted a plain STEP file. Can't do MBD if no one wants MBD and the software doesn't support all of the required characteristics to run full factory planning from the model.
On a large job for the IDF corporate had bid on a build-to-print with design responsibility with an Israeli company. Unfortunately the Israeli company had an on site factory crew who figured out how to build things and so no welding information was on the drawings. What was worse is build-to-print meant no budget was allowed to engineering so our manufacturing did what the Israeli company did and let the welders do whatever they wanted and documented none of it; this is a problem when TACOM, the stand-in for the IDF, wants to ensure the production units are just like the qualifying prototype. Can't make them the same if no one knows what that should be.
So I got the job of documenting the 20-30 pounds of weld material - all represented in the assembly model to show the expected extent of the weld. It also uncovered a bunch of areas the Israeli model diverged from the Israeli drawing. They used UG, so we imported STEP. Likewise they did not include materials; apparently there was a misunderstanding and they had the idea that the material spec (ASTM) specified material, not knowing that some ASTMs cover dozens of alloys, from soft to nearly armor plate, but they failed to record what specific ones were used on the prototype.
At the end each part had a model material assigned, the correct ASTM spec, and the weldment had welds modeled at the correct size an in the order the welds were required. The user for that was drafting as the IDF wanted PDFs and the drafters were to add annotation to the drawings to match the depicted welds.
The stress group did sometimes use the native models, but for many they too would STEP out the geometry, cut back features, simplify shells to midplanes.
On the flip side; I've never seen a factory model that includes a statistical model of ability to conform to requirements. Like, for CNC mill XYZ what is the distribution for variation from true position for typical hole diameter statistical bins, and what is the composite range across the factory? Not sure there is software for the purpose of pre-qualifying a factory to produce a part.
So, worked very hard in that direction, but procurement, the factory, and the customer all had separate leadership and couldn't or wouldn't use MBD, just STEP extraction for fabrication/CNC toolpath.
