Questions we answerAI and automation

Should I automate this work?

If it is repeatable, frequent, low-judgement and written down, quite possibly. If any of those is missing, fix that first.

Start from this question

Free · a few minutes · starts from “I think AI could help

The short answer

Automate when the steps are the same each time, the process is written down, it comes round often enough for the setup to pay back, and someone can say what happens when it goes wrong. Do not automate work that lives in people's heads — you will automate the mess. Do not automate work that comes round twice a year, because the setup costs more than the work. And check the system you already pay for first: its own automation is already governed, logged and supported, which a new stack would not be.

Automation pays back on volume and consistency. It fails on work that is different every time, on processes nobody has written down, and on decisions that need a person.

The diagnostic checks each of those before it says an agent is credible, and it will say plainly when simpler automation or a person with better tools is the better answer.

Is it really this?

Worth checking before you spend anything. These are the signals that tend to tell the two apart.

Signs it is

  • The steps are the same every time and someone could write them down in an afternoon.
  • It comes round often enough for the saving to be worth the build.
  • The inputs already live in a system rather than arriving as emails and attachments.

Signs it's something else

  • Most of the time goes on exceptions rather than the standard path.
  • There is a judgement call partway through that changes what happens next.
  • It is rare enough that maintaining the automation would cost more than doing the work.

What's usually behind it

  • A process that has never been simplified, only repeated.
  • Copying between systems that could talk to each other.
  • Work that is repetitive for a person but still needs a judgement call at the end.

What people usually get wrong

  • Automating before simplifying, so the wrong thing happens faster and is harder to change.
  • Not deciding who owns the automation on the day it breaks.
  • Counting a saving in hours that nobody reclaims for anything in particular.

The options, and when each one fits

In no particular order. Which of these fits depends on your situation, not on the question you typed.

  • Stop doing it

    The step exists out of habit and nobody would miss the output.

  • Connect the systems

    The work is moving the same information between tools that could exchange it.

  • AI assistance with a person in charge

    The routine part can be sped up but the decision should stay with someone.

  • Automate the workflow

    The steps are documented, frequent, low-judgement, and there is an owner for when it fails.

When doing nothing is the right call

If the work is infrequent, well understood and nobody is straining under it, leaving it alone is a reasonable answer. Automation has a running cost — it has to be watched, corrected and updated — and small tasks do not always repay it.

What the diagnostic asks

  • How repeatable the work is and how much judgement it needs.
  • How often it comes up and what happens if it goes wrong.
  • How sensitive the information is and whether there is already a system for it.

“I don't know” is always an answer. Nothing is assumed on your behalf.

The routes it weighs

Every diagnostic weighs all ten routes. These are the ones that tend to matter for this question — a description of the rules, not a promise about your answer.

  • AI Agent

    Delegate a defined, repeatable workflow to software inside boundaries you set, with human oversight.

  • AI Assist

    Keep a person in charge and give them AI for drafting, summarising, checking or analysis.

  • Process improvement

    Change how the work flows — remove steps, name an owner, fix the hand-offs.

What changes the answer

  • A process that lives in people's heads has to be written down before it can be automated.
  • High stakes mean a person approves what the agent does, at least at first.

Get your own read

A handful of short questions, then the strongest route for your situation, what's also credible, what we still don't know and what would change our view. No account, no setup.

Related questions