Build the house yourself or buy it turnkey?
Sounds like a simple decision. It isn't. Not for houses, and not for software.
AI has made the first step of building easier. What used to take months now takes hours.
But: possible isn't the same as sensible.
Because AI only lowered one barrier, the barrier to building. Not the barrier to maintaining. Not to integrating. Not the question: who looks after this in 6 months?
How we frame the question
At AssetOS we ask the same make-or-buy question constantly. And we don't always answer it the same way.
Two examples from the last few months:
We built: our own AI agents for error analysis in our AWS environment.
Why? We already had agent infrastructure for code reviews, the incremental step was small. Datadog with Watchdog as a full observability platform would have been overkill. The effort was manageable. The benefit direct.
We bought: HubSpot, for sales and marketing.
Why? Because as CTO I'd otherwise burn 60% of my time to be worse than a product improved daily by thousands of teams. The opportunity cost would have been brutal.
The principle
Build where your competitive advantage lies. Buy everything else.
For a tech startup, that's sometimes the dev stack itself.
For a real estate investment team?
The competitive advantage is deal judgment. Relationships. Speed of decision. Not the ability to build and operate software.
Build where you make the difference. Buy everything else.
Curious where your teams draw this line, and whether AI has shifted it.