How we calculate ROI on an automation project

Every proposal we send includes an expected payback period. Here's exactly how we calculate it, so you can check our numbers or run them yourself.

Step 1: Measure the baseline

We count how often the task happens and time it in real conditions, not the best case. Then we apply a fully loaded hourly cost: salary plus taxes, benefits and overhead. For many roles that comes to 1.25 to 1.4 times the base wage.

Baseline cost = frequency × minutes per task × loaded hourly cost

Step 2: Use a realistic automation rate

Very few processes are 100% automatable. Documents arrive in odd formats, customers ask unusual questions, systems go down. We typically model 50–80% of volume handled automatically, with the rest flowing to a person, who is still faster than before because the system has done the prep work.

Step 3: Count all the costs

The build fee is the obvious one. These are the ones that get forgotten:

• Running costs: AI model usage, automation platform fees and hosting
• Review time: the minutes your team still spends checking exceptions
• Maintenance: systems and APIs change, so budget a few hours a month

Step 4: Compare

Monthly savings = (baseline hours × automation rate × hourly cost) − running costs − review time

Payback period = build cost ÷ monthly savings

A healthy project pays back within three to nine months. If our estimate comes out longer than twelve, we'll usually recommend a smaller first step, or tell you it isn't worth automating yet.

Then measure again

After launch we compare actual hours against the baseline. That number, not our estimate, is the one that matters, and it's the one we report back to you.

Want to try it with your own numbers? Our savings calculator is a good place to start.