Automating a Slow Manual Workflow: Sizing the Cost, Rolling Out and Measuring Impact

Describe a process you made faster or more reliable through automation. What was wrong with it, what did you automate, and what changed? ```hint Show the cost first Think about how you knew this process was worth fixing, and what you measured about it before anything was automated. ``` ### Clarifying Questions - Does it count if an engineer built the automation and my role was to identify the problem, define the rules and drive adoption? - Should the process be part of the product customers use, or is an internal team process such as release checks, feedback triage or reporting a fair example? - Are you interested in AI-based automation specifically, or does any approach count, including scripts, workflow tools or rules engines? ### What a Strong Answer Covers - How the candidate found the process and sized its cost (hours, errors, delays, who was affected) before choosing a fix - Why automation was the right remedy rather than removing or simplifying the step, and which parts were deliberately left to people - How the change was rolled out: a pilot, handling of exceptions, and bringing along the people whose work changed - Before-and-after measurement of the result, including any new failure modes the automation introduced - The candidate's specific role and what they would do differently ### Follow-up Questions - After launch, what case did the automation get wrong, and how did you find out? - How did the people who used to do this by hand react, and what did you do about it? - How did you decide the automation was worth its build and maintenance cost?