Thanks for such a detailed question! We're focused on material quantity estimates right now. We're using pretty generic averages for costing as our primary users (suppliers) have a way better finger on the pulse of the market and typically change the unit price anyways. We'll work on better cost accuracy though, and are looking to integrate with RSMeans or building our own scrapers.
For the second question, it really is most accurate on "fully standardized" blueprints due to our training distribution. Will work on improving that as well!
I realize you're opening yourself up to criticism if you answer this truthfully, but since suppliers are your primary user (and therefore paying customers I assume?) your pitch is "supplier, you can do more with less (or no!) people when specifying material quantities, and in fractions of a second as opposed to the minutes/hours it takes today!" So, ultimately, if the suppliers fully trusted your solution they would need zero personnel determining quantities? And since your solution's annual licensing cost would likely be a fraction of the price of even a single individual that'd be pretty compelling to the supplier.
Best of luck with the business (and with getting to know the corp dev people at Autodesk/Procore/etc.--sorry, couldn't help myself!).
> f the suppliers fully trusted your solution they would need zero personnel determining quantities?
You'd still need people to check what actually got installed, so that you can bill for it. Like, there's only so much you can determine off the plans.
And what happens if (when) the plans are wrong or impractical?
My Dad worked in construction for his career, and I did briefly, and there's generally a lot of stuff that needs to be figured out on site due to physical or logistical constraints.
Sounds like this is just for homebuilding though, which is a much easier problem.
It seems like, with sufficient users entering data (including previous sales/purchases if desired), you could begin to train models to take into account local/regional factors etc. do you have any long term plans to collect user data in such a manner?
Separately, it seems like it would be incredibly useful to use your models in various embodied carbon estimation tooling and other decarbonization research streams. Have you thought about partnering with any academic researchers on this? If you are interested, let me know, as I can definitely connect you with a bunch of researchers who would be interested!
We don't really have plans on using user-data to train our models, since it's too much of a chicken-and-egg problem. Carbon estimating is something we had some clients talk about -- it uses the same material quantity estimation we're already doing, and is something we're really interested in (my heart is in impact-driven work). Would love the connections!
For the second question, it really is most accurate on "fully standardized" blueprints due to our training distribution. Will work on improving that as well!