Skip to main content

Questions we answerOutsourcing and external support

Should I hire an AI consultant or an automation agency?

Neither, until you can name the process and what better looks like. A consultant sells a decision; an agency sells a build. Buying the build first is how you end up with workflows nobody can maintain.

Start from this question

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

The short answer

Neither, until you can name the process and say what better would look like. A consultant sells you a decision; an agency sells you a build, and buying the second without the first is the reliable way to end up with workflows nobody internally can operate. If you already know the problem and just need it built, scope a small paid pilot with a named internal owner and a handover you could actually run — and treat the handover as the deliverable, not the workflow.

Every page answering this question is published by somebody selling one or both services, and the decision-guide format is being used as a qualification step: both branches end at a contact form.

The distinction worth making is one that neither seller has a reason to draw. Paying for a decision and paying for a build are different purchases with different failure modes, and they are routinely sold as a single engagement — which is how a discovery workshop turns into a slide deck, or a build turns into a dependency.

Is it really this?

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

Signs it is

  • You can state the process, the current cost and what better looks like, in writing.
  • There is somebody internal who will own the result and has the time to.
  • The scope is small enough to be a pilot you could walk away from.

Signs it's something else

  • The brief is AI rather than a problem, which means you would be paying somebody to find one.
  • Nobody internal will own it, so whatever is built becomes a permanent external dependency.
  • The process changes every time, which is a process question before it is a technology one.

What's usually behind it

  • A sense that AI should be doing something here, without a specific process in mind.
  • A previous attempt that produced something nobody uses.
  • Budget approved for AI in the abstract, and a search for somewhere to point it.

What people usually get wrong

  • Buying a build before a decision, and discovering the decision was the hard part.
  • Accepting a handover that consists of access rather than understanding.
  • Treating partner badges and certifications as evidence of quality. They are evidence of a commercial relationship.

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.

  • Neither yet

    The process is not named and the outcome is not defined. Anything bought now is a discovery exercise you could do yourself.

  • Buy a decision

    You have candidate problems and no way to judge which is worth doing, or whether any is.

  • Buy a build

    The decision is made, the owner is named, and the handover is written into the scope.

  • Do it internally

    The problem is bounded and somebody inside has the capability and the time to own it.

When doing nothing is the right call

If you cannot yet name the process, doing nothing for a month while you do is not a delay — it is the work. It is also the cheapest possible version of the discovery engagement you would otherwise be buying.

What the diagnostic asks

  • Where you think AI could help and how repeatable that work is.
  • How much judgement it needs and how often it comes round.
  • What a mistake costs and how sensitive the information is.
  • Whether software already exists for it, and who owns the process today.

“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.

  • Specialist support

    Bring in a freelancer, consultant or specialist for a defined piece of work.

  • Outsource

    Hand a defined, ongoing scope of work to an external provider.

  • Do nothing yet

    Record the issue, set a revisit date, and decide what would change your mind.

  • Process improvement

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

  • AI Assist

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

  • Develop

    Build the capability in someone already in the team, with time and support.

  • AI Agent

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

What changes the answer

  • Being able to name the process and the target outcome yourself removes most of the case for buying a decision.
  • Having nobody internal who could own the result afterwards is the strongest argument for waiting.
  • Existing software that already covers it changes the answer entirely, and a seller of builds has little reason to look.

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