Automation Opportunity Finder

Why this question matters

Automation is most useful when a recurring task has a clear purpose, a recognisable process and an appropriate place for human judgement. A large amount of manual work does not automatically make a good automation candidate. This finder compares up to eight processes using effort, repetition, clarity and judgement. It helps choose a bounded pilot and highlights work that should retain human control. The displayed time release is a scenario, not a promised saving.

Describe each process narrowly enough to measure. Copying approved order information into a dispatch record is more useful than automating operations. Enter occurrences per week, minutes per occurrence and the number of people involved. Be clear whether the time represents each person working on the occurrence, because the formula multiplies by people. If one person hands work to another for different durations, use a defensible average or separate the activities.

Rate process clarity according to how consistently the steps and exceptions are understood. Rate judgement separately: one represents routine rules, while five represents essential human judgement or relationship care. Report the percentage of occurrences with an error using a consistent definition. The error input provides context but does not directly increase the time release estimate. A task with frequent errors may need a process correction before automation is considered.

Understand the method

1. Annual effort hours = times per week × minutes per occurrence ÷ 60 × people × 52.
2. Illustrative release fraction = 0.75 × process clarity fraction × (1 − judgement fraction).
3. Potential released hours = annual effort × illustrative release fraction.

Annual effort equals occurrences per week multiplied by minutes per occurrence, divided by sixty, multiplied by people and by fifty two weeks. The illustrative release fraction is 0.75 multiplied by the normalised clarity rating and by one minus the normalised judgement rating. Ratings are converted so that one maps to zero and five maps to one. Potential released hours equal annual effort multiplied by that fraction.

The seventy five per cent ceiling is an explicit modelling choice, not an observed automation saving. Complete clarity and minimal judgement reach that ceiling. Essential judgement produces zero estimated release under this model, even where parts of the surrounding administration could potentially be assisted. Low clarity also suppresses the estimate. Read the result as a prioritisation aid that deliberately favours clear, routine work.

Select one process and test a small sample with known expected outputs. Measure current time, error patterns and the effort needed to correct exceptions. Then compare the assisted version using the same work. Include setup, review and maintenance effort when assessing value. Keep a manual route and a person responsible for stopping the pilot if output quality or information handling becomes unacceptable. Released time becomes useful only when the business decides what to do with it.

Keep the result in perspective

This is a transparent planning calculation or self assessment, not a sector benchmark, professional valuation or a prediction of an outcome.

The 75% ceiling is an explicit scenario assumption, not observed savings. Errors are reported separately and not monetised. Time released is not automatically payroll saved. Validate security, consent and exceptions before a pilot.

The model assumes the entered weekly pattern continues for fifty two weeks. It does not account for seasonal demand, holidays, different job sizes or time shared between people. It does not price software, integration, maintenance or training. Errors are displayed separately and are not monetised. Avoid interpreting the annual hours as payroll savings, because salaries do not automatically fall when a task takes less time.

The finder does not inspect data security, consent, system access or legal obligations. An apparently routine process can still involve sensitive information or consequential decisions. Review the actual information flow, permissions, external services and recovery route before a pilot. Do not put credentials, personal records or confidential customer content into an experimental automation merely because the time calculation looks attractive.

Read the full limitations or explore how Ayodeji approaches this work.