Think Build Implement Repeat
London, UK +44 7367 067226
WhatsApp FOLLOW f in X
AI Integration

The Work That Continues After Launch

Last updated:

It is not like maintaining a website

Conventional software keeps doing what it did. AI integrations degrade quietly as your inputs change, your content ages and providers update their models.

That is why the maintenance figure is higher than people expect and why skipping it produces a system nobody trusts within a year.

The monthly work

  1. Review the evaluation results and the correction rate trend
  2. Look at what reviewers corrected most, and why
  3. Check cost per item against last month
  4. Confirm the index is current and nothing stale has crept back
  5. Adjust thresholds if the evidence supports it
Two to four hours a month, done consistently, is the difference between a system that improves and one that decays. It is not difficult work; it just has to happen.

The occasional work

  • Provider model changes — rerun evaluation, adjust if needed
  • New document types or customers — extend the evaluation set
  • Credential rotation, before expiry rather than after
  • Annual review of provider terms and pricing

Who does it

Someone in your business owns quality — reading the numbers and deciding. Someone technical makes the changes, whether that is your team or ours.

Both roles need to be named. Unowned systems are the ones that quietly stop being useful.

What it costs

SystemAnnual maintenance
Simple integration, stable inputs15–20% of build
Document processing, varied inputs25–30%
Customer-facing assistant25–35%

Budget it explicitly at the start. Discovering it in year two, unbudgeted, is how good systems get switched off.

Frequently asked questions

Can our own team do this?

Yes, with the handover session and the runbook. The judgement part is business knowledge more than technical skill.

What if we do nothing?

It works, then gradually works less well, and someone eventually declares it a failure. The decay is slow enough that nobody notices the cause.

Is retained support worth it?

Where quality has commercial consequences, generally yes. For a stable internal tool, ad hoc is often sufficient.

Does maintenance decrease over time?

Somewhat after the first six months, as edge cases are absorbed. It does not go to zero, because your business keeps changing.

Keep reading

Have an integration nobody has looked at in six months?

It is probably worse than it was. A review is a couple of days and usually finds two easy improvements.

Book a free 30-minute call Get a project estimate WhatsApp us

Related services

What we build for problems like this one

AI IntegrationEnterprise AIAI Agents