The message usually comes on a Friday: "Our architect is leaving." Suddenly the model that's been quietly running in the background — the one nobody else fully understands — needs a new owner, and hiring one takes months.
Here's what actually breaks first, and what to do about it.
What stops immediately
Not the model itself — models don't expire. What stops is the judgment layer: someone deciding whether a proposed change fits the target architecture, someone who remembers why a decision was made eighteen months ago, someone who can sit in a steering committee and say "no, that breaks three things downstream."
Projects don't wait for a hire. They keep moving, and without that judgment layer, they either stall waiting for sign-off, or worse, they proceed without it and the model quietly stops being true.
The short-term fix isn't a full-time replacement
Most organizations over-scope this. You don't need someone embedded five days a week from day one — you need someone who can keep the model current, keep architecture reviews running, and make sure nothing gets decided in a vacuum, while you take the time to hire properly rather than rushing it.
That's interim cover: a few focused days a month, structured around your existing review cadence, so the gap doesn't compound while you look for the right permanent person.
What to have ready before the gap starts
- A short handover document — even an incomplete one is better than nothing. Ask the departing architect for 90 minutes before they go.
- A list of in-flight decisions waiting on architecture sign-off, so nothing silently stalls.
- Access to whatever tool holds the model — Ardoq or otherwise — confirmed and working before the last day, not after.
The goal isn't to freeze everything until you hire. It's to keep the practice breathing — reviews still happening, the model still trustworthy — so that whoever you eventually bring in inherits something usable, not a six-month backlog.