How to estimate automation ROI before you build

Start with monthly process volume, multiply by average staff time and cost, then add the cost of errors and opportunities lost to slow replies. That is the current-state cost.

Model a realistic scenario

Do not assume 100% automation. Use a conservative share of work completed, then add human review, platform usage and maintenance.

Measure after launch

Track cycle time, conversion, human handoffs and error rate. If those numbers do not improve, fix the workflow before adding more features.

Released time is not always cash saved

Measure request volume, active handling time and correction effort before automation. Separate customer waiting from employee work. Released time creates capacity, but does not automatically reduce spending when staffing costs remain unchanged. Identify whether that capacity will serve more requests, reduce overtime or improve follow-up. This prevents a productivity estimate from becoming a misleading cash-savings claim.

An illustrative calculation

Assume 1,000 monthly requests, four minutes released per request and an internal labor value of EGP 90 per hour: 1,000 × 4 ÷ 60 × 90 = EGP 6,000. Replace these hypothetical inputs with observed figures.

Test sensitivity and avoid double counting

At a hypothetical EGP 18,000 setup cost, value payback is six months if the assumptions hold. Halving requests reduces time value to EGP 3,000; unchanged EGP 3,000 costs leave nothing for setup recovery. Include training and maintenance in scenarios. Do not simply add all time value to sales generated using that same time, or attribute every sales increase to automation when advertising and prices also changed.