The failure that makes no noise
Automations rarely crash. They stop producing, or produce empty results, and nobody notices for weeks because the dashboard is green and no error was raised.
The most under-implemented alert in business automation is “this produced nothing today and it usually produces forty”. We build that one first.
What we monitor
- Volume outside the expected range for that day of the week
- Zero output where output is normal
- Field completeness, so a format change upstream is caught the same day
- Third-party integration health
- Quarantine queue growth, with an owner who reviews it
Warranty, then maintenance
Ninety days of defect fixing after launch is included — behaviour that does not match what was agreed. After that, maintenance is optional and priced at 15–25% of build cost annually depending on how many third-party systems are involved.
| Build cost | Annual maintenance | What it covers |
|---|---|---|
| £10,000 | £1,500–£2,500 | Monitoring, dependency updates, defect fixes |
| £25,000 | £3,750–£6,250 | The above plus integration repairs |
| £50,000 | £7,500–£12,500 | The above plus a small change allowance |
What counts as maintenance and what does not
Maintenance keeps it working as the world changes: security patches, dependency updates, repairs when a supplier alters their API. Development changes what it does.
We write the boundary into the agreement with examples, because almost every maintenance dispute in this industry comes down to that sentence being absent.
The annual review
- Does this still reflect how you actually work?
- Is the output still used by someone?
- Which exceptions are being handled manually that could now be automated?
- Has anything changed upstream that we have not accounted for?
- Should this still exist at all?
We have recommended switching automations off. A system that no longer serves a purpose still consumes maintenance and still occasionally fails.