If I had a dollar for every time I get pitched a new start up proptech operations platform, I probably wouldn't need to be fundraising for our IOS deals.
But that's not where we believe the future lies.
Two years into building AI tools at Westlake Realty Group, we don't want to replace Yardi. Instead, we have built on top of it.
Here is why.
Your PM software is not just a tool. It is a record of everything your company has learned about every tenant, every lease, every vendor, and every property for over a decade of operations (we were with MRI before then). Work order history. Payment patterns. Lease abstracts. CAM reconciliations. Vendor performance. Over a decade of data and knowledge.
When you rip and replace, you lose the context. You spend a year migrating data, retraining your team, and rebuilding workflows inside someone else's product architecture. You end up with the same data in a new container.
We kept Yardi where it is. Then we built Claude skills that sit on top of it and read what it knows.
A few examples from our PM operation (and I won't even make you comment AI to see these):
1. Critical date tracking. Our skill reads the lease data in Yardi and surfaces what needs attention this week: renewal windows, notice deadlines, option periods. The data was always there. The skill tells Kelly Chang Levine what to do with it.
2. Work order routing. The skill reads our vendor list, the work order type, and the property history and routes it to the appropriate person. Scope decisions stay with Kelly. The skill handles the lookup logic so she does not have to. The associated agent sends Kelly a digest every day at 4pm.
None of these exist because we bought a platform. They exist because we know our operation well enough to describe it precisely.
The data was always in Yardi. What we built was the interpretation layer.
This is what proptech vendors will not tell you: the software is not the constraint. Knowing what to ask is.
Are you building the intelligence layer, or shopping for a new container?
Discussion about this post
No posts
