SurfingBear ToolsSurfingBearTools
Skip to content

Home › Blog › Running an AX assessment yourself

How to run an AX assessment without hiring consultants

Before asking for a consulting quote, you can do a substantial part of the assessment yourself. Here it is in four steps, with the point where you genuinely need outside help marked.

4 steps In-house Where outside help starts

The short answer

An AX assessment is not fundamentally a special methodology. It is writing down, exhaustively, what people currently repeat, and judging which of it can move to a machine. About 70% of that work can be done by your own staff.

The point where outside help becomes necessary is specific: when the judgement needs technical grounding, and when departments disagree about priority. Up to there, doing it yourself is faster and costs nothing.

Step 1 — Inventory the work (do this yourself)

Ask each department to write down, in order, what they actually did last week. Not the job description — the actual actions. That difference matters. The job description says "market analysis"; the reality is usually "copy numbers from three sites into a spreadsheet and build a chart". The second one is the automation target.

Record three things alongside each task: how often it happens, how many minutes it takes each time, and who receives the output. Without those three you cannot prioritise in the next step.

Step 2 — Classify by repeatability (do this yourself)

Fully repetitive — The processing sequence is always the same even when the input differs. First priority for automation.
Conditionally repetitive — It branches into a few cases, but each case is handled a fixed way. Second priority.
Requires judgement — A person interprets something every time. AI can assist, but this is not a candidate for full automation.
One-off — Does not repeat. Excluded from automation.

Step 3 — Estimate the benefit (yourself, but show your working)

For each task, calculate the annual hours: frequency × minutes per occurrence × number of people × 52 weeks. Then assume an automation rate conservatively. Starting at 50% is safer. Even if you end up achieving 70–80%, setting the bar low and exceeding it is what earns trust inside an organisation.

The common mistake at this step is converting saved hours straight into headcount savings in the report. In practice headcount does not fall — that time moves to other work. Reporting saved hours as "time recovered", together with what will be done with it, has a much better chance of approval.

Step 4 — Set priorities (where outside help pays)

Plot impact against difficulty and start with high impact, low difficulty. You can get this far in-house.

What needs outside help is the difficulty judgement. How technically hard a given task is, is difficult to estimate without build experience, and estimating it wrong means your first project turns out to be the hardest one — and fails. Departmental disagreement about priority is the other case; internal staff rarely have the standing to arbitrate it. Those two things are the real value of a formal assessment.

When a self-assessment is enough, and when it is not

Self-assessment is enough

One or two departments, fewer than ten automation candidates, and a technical person in-house. On those conditions, going straight from self-assessment to a pilot build is faster.

You need outside help

Five or more departments, tangled system integrations, or a priority decision that needs agreement between departments. That is the point where the cost of coordination exceeds the cost of assessment.

Frequently asked questions

Can we use our self-assessment output directly in a consulting engagement?

You can, and you should. Finishing steps 1–3 in-house shortens both the duration and the scope. With a task inventory and time estimates in hand, the engagement starts at step 4.

What if staff overstate or understate their work?

A common problem. People understate when they fear automation means job cuts. Sharing the policy up front — that recovered time moves to other work rather than reducing headcount — measurably improves accuracy.

How long does it take?

For two or three departments, about a week to inventory and a few days to classify and estimate. With more departments, interview scheduling becomes the bottleneck, so two to three weeks is realistic.

Can we skip the self-assessment and go straight to a formal one?

Yes. But doing the self-assessment first means the same engagement duration covers more ground. Without prior material, a large share of the engagement goes into establishing the current state.

Steps 1–3 yourself, step 4 together

Bring us your self-assessment

With a task inventory and time estimates in hand, a consultation starts at the prioritisation step. Halfway through is fine too — tell us where you are and we will review it and reply.

Enquire about an assessment