You've outgrown the system you built. Now what?

The first system a growing business runs on is almost never designed. It accretes. An Airtable base someone set up in a weekend. A Monday board for the jobs. A spreadsheet for stock because the base didn't have a stock table yet. A few automations that one person understands. It worked, because the business was small enough that a person could hold the gaps in their head.

Then the business grows, and the gaps stop fitting in anyone's head. The symptoms are the same everywhere: the same job typed into two places, a quote that doesn't match the invoice, a weekly report that takes a morning of asking around, and one office person who can't take a holiday because the process lives in them.

The two bad options

Owners usually see two choices. Live with it, which means hiring someone to do the re-keying and accepting that growth makes the office worse. Or start over, which means a big quote, months of nothing, and losing the parts of the old system that actually worked. Both are expensive, and the second one is where most burned-by-a-builder stories start.

The third option: keep what's sound, upgrade what isn't

Almost no outgrown system needs replacing. What it needs is a map. Which tables are the real source of truth. Which ones are copies. Where data is typed twice. Which automations run and which ones silently stopped in March. Once that map exists, the upgrade is a short list of specific jobs, in order, and most of the original system survives.

• Clean the data first. Every system built on dirty records fails on day one.
• Connect what's already there before adding anything new.
• Add AI only where a person used to read, judge or draft. Half the work is plain automation.
• Ship one piece a week and let the people who use it check it on Friday.

The test of whether it worked isn't the demo. It's whether, a month later, the office manager runs the system without calling anyone.