We replaced HubSpot, then the real work started

When I posted that we replaced HubSpot with our own software, the build is what got people's attention. That makes sense, because it's the fun part. But if you're thinking about doing the same thing, the build isn't the part I'd worry about most.
We built our own CRM. Our CFO built most of our finance tool herself. We built a tool to prep for business reviews with our customers, and it has saved us somewhere between 500 and 1,000 hours. None of that came from a big software team. It came from people who know the work, using AI build tools.
So yes, I'm a believer. I use these tools every day and I love it. But building was the easy part, and I think a lot of owners are about to find that out on their own.
The build is the fun part
You describe what you want, a screen comes up, you click around, you change your mind, you change it again, and in a few days you have something that would have taken months not that long ago. It feels like magic. It's also easy to stop right there and call it done.
Then the next week comes. Someone finds a bug, someone wants a new field, someone asks why the report doesn't match what they see in the other system, and someone new starts and needs a login. None of that is building, but all of it is now part of the job.
Who owns it once it's live?
The first question I'd ask is who owns the thing after it's built. The person who built it may not be the right person to own it, and if nobody owns it, it starts to drift. Someone has to decide what gets changed, what gets fixed first, and what gets a no. With a tool you buy, the vendor makes a lot of those calls for you, whether you like them or not. When you build it yourself, those calls come home to you.
That sounds simple, but it's where I see the most risk. A tool that one person built on a Saturday can end up running a real part of the business, and the rest of the team may not even know how it works. That's okay for a test, but it gets risky once your sales or finance work starts to lean on it.
Who keeps it safe?
When you buy software, you are also buying part of someone else's security team, and when you build your own, that job comes back to you. Who can log in, what data it can see, what happens when someone leaves, where the backups live, and how it gets patched. Those questions don't go away because the build was quick.
We take this part seriously because it's what we do for our clients every day. When we started building with a team of AI agents, one of those agents had a single job, to try to break what the others built, and it has caught gaps we would have missed. That helps a lot. You still need a person who owns the security of what you built and checks it on a schedule, the same way you would for any other system in the business. If that isn't something your team has time for, that's what managed security is for.
Who keeps it useful?
A tool nobody uses isn't worth much, no matter how fast you built it. A CRM is only as good as what goes into it. If people don't put their notes in, or the old data comes over messy, you end up with the same headache you had before, just with your own name on it. That's as much a people and process problem as a software problem, and the build tools won't solve it for you.
Useful also means it keeps up. The work changes, the team changes, and the tool has to change with it. Changing your own software is a lot easier than it used to be, which is great, but somebody still has to be paying attention to know what needs to change.
What I'd ask before you build
If I was advising a client who wants to build instead of buy, I wouldn't talk them out of it. I'd ask a few plain questions first:
- Who owns it once it's live
- Who keeps it secure, and how would you know if something went wrong
- What happens when the person who built it is out for two weeks, or leaves
- How will you know people are actually using it
If you have good answers, go build it. If you don't, you can still build it. Just go in with your eyes open, and maybe start with something small that won't hurt much if it breaks.
Why we went first
That's a big part of why we did this for ourselves first. We're our own first customer. We hit the bumps on our own tools so our clients don't have to learn it the hard way. When we help a client with purpose-built apps, we spend as much time on who owns it, who secures it, and how it gets used as we do on the build. And once it's live, someone has to watch it, update it, and keep the AI parts doing what they should, which is where managed AI comes in.
I still think building your own software is one of the best things AI has made possible for a business like ours, and for a lot of our clients too. I just want people to know what they're signing up for.
If you built your own tool this year, who's going to keep it running a year from now?

