Home Services My Story Work Press The Desk Start Here Book a Call
The Desk  ·  Measurement

How to Calculate
Automation ROI

Most companies measure the wrong thing after automation. Here is what to measure, when to measure it, and how to build the baseline before you start.

8 min read June 2026 Priyankka Wadhwa
Scroll

The Number Nobody Measures Before They Start

Most companies measure ROI wrong. They look at what they saved after the build. The right question is what it cost before.

Without a baseline, there is no number. There is a feeling. "It feels faster" is not ROI. "We were spending 18 hours a week on this process and now we spend 2" is ROI. The difference between those two statements is the baseline measurement.

Most teams skip the baseline because they start thinking about ROI after the build, not before. By then the before is a memory, not a measurement. The number gets estimated. The estimate is always optimistic. The ROI claim becomes unverifiable.

The only way to calculate automation ROI accurately is to measure before the build starts.

Why the Baseline Is Not Optional

A baseline is not a bureaucratic step. It is the only thing that makes the ROI calculation real.

When the baseline is missing, three things happen. The team cannot prove the automation worked. The sponsor of the project cannot justify the cost to leadership. The next automation proposal gets more resistance because the first one produced only a feeling, not a number.

The baseline takes two weeks to measure. For most processes, it takes about four hours to set up: an hour to identify what to track, an hour to agree on what counts, and two weeks for the data to arrive. Then the build starts.

The Four Things to Measure

  1. 01 Manual hours per week. Count every person who touches the process. The person who runs it, the person who reviews it, the person who approves it, and the person who handles exceptions. Add them up. This is the weekly time cost of the process.

    Most teams undercount here. Managers count the person who owns the task. They miss the two people who spend 20 minutes each week answering questions about it.
  2. 02 Error rate. How often does the manual process produce an incorrect output? Track for two weeks before the build. Define what counts as an error before you start measuring, not after, when the definition becomes conveniently narrow.

    Error rate matters because automation typically reduces it to near zero. If you did not measure it before, you cannot claim that result.
  3. 03 Response time. How long from the trigger to the completed output? An invoice received to an invoice logged. A client query received to a first response sent. A report requested to a report delivered.

    Response time is often more visible to clients than the team realises. Measuring it before and after produces a number the business can feel and the client can notice.
  4. 04 Monthly process cost. Hours per week × 4 × average hourly rate of everyone involved, plus any tool costs, plus an estimate of rework cost when errors are caught and corrected. This is what the process costs the business per month.

    This number is often a surprise. Processes that feel like background work turn out to cost £8,000 to £15,000 a month in staff time alone. The surprise is useful. It changes the conversation about whether to automate from "can we afford to?" to "can we afford not to?"
If you cannot state what this process costs per month before you build, you are not ready to measure ROI after.

What Happens at 30 Days

At 30 days, compare against the baseline. Four numbers: hours then, hours now. Error rate then, error rate now. Response time then, response time now. Monthly cost then, monthly cost now.

Most teams see 60 to 70 percent reduction in time spent on the automated process in the first month. The number varies based on how manual the starting point was. Export houses and family businesses that were running entirely on spreadsheets and WhatsApp groups tend to see the largest early drops.

What the team does not expect is how the recovered time feels. For the first two weeks, the hours exist but do not feel fully available. The team keeps expecting the process to break. They check it more often than they need to. Then they stop. Then the time is genuinely theirs.

The 60-Day Signal

At 60 days, something shifts in the language. The team stops asking "can you automate this?" and starts asking "can we automate this?"

That single word change is the most reliable signal that the implementation has held. The team has moved from user to owner. The question is no longer permission-seeking. It is capability-claiming.

At 60 days, the team also starts pointing at the next process. The one they point to first is almost always right. It is the process they have complained about the longest. It is usually the one that involves the most manual data movement or the most approval steps with no clear logic.

"At the 60-day review, the team had already started automating their own onboarding checklist. Nobody asked them to. They had just watched what was possible and started applying it. That is the signal you want, not the ROI number."

What Compounds at 90 Days

At 90 days, isolated automations start connecting. The output of one pipeline feeds the input of another. The data that was extracted in step one becomes the trigger for step two, which feeds step three.

This is where ROI compounds. The individual automations each have their own savings. The connected sequence produces savings that multiply.

In one engagement, a 17-person SaaS company in the United States, three workflows were built over 90 days. The first handled client onboarding data. The second connected that data to project setup. The third tracked project milestones and sent automated status updates. None of the three workflows was individually remarkable. Together they eliminated 14 manual spreadsheets and six hours of daily data entry. The result: $340,000 saved in 90 days, measured against the baseline cost of the three manual processes they replaced.

Not from one automation. From the sequence working as designed.

The Calculation

Simple version:
Before monthly cost − After monthly cost = Monthly ROI

With implementation cost factored in:
Monthly ROI × 12 = Annual ROI
Annual ROI ÷ Implementation cost = Payback period in years (multiply by 12 for months)

For most desk-level automations, the payback period is between one and four months. For larger team rollouts, three to eight months.

The calculation only works if the before number was measured, not estimated. Which is why the baseline comes first.

The full implementation methodology, how the audit phase sets up the measurement framework before the build begins, is in the implementation guide.

A Note on Indian MSMEs and Family Businesses

For solopreneurs, founders, and family businesses in India, the ROI calculation often includes a dimension that does not appear in the formula above: the founder's personal time.

Most small business owners in India are the process. They are the one who sends the follow-up, approves the order, handles the exception, and reviews the report. When those tasks are automated, the savings are not a reduction in staff cost, they are a return of founder time.

That time has no line in the calculation. But it is almost always the first thing a founder mentions at the 30-day review.

ROI Measurement at Each Stage

Stage What to Measure
Before build Baseline: manual hours/week, error rate, response time, monthly process cost
Day 30 Compare all four metrics against baseline. Typical result: 60–70% time reduction
Day 60 Ownership shift: "can we automate this?" Team points to next process
Day 90 Compounding: isolated automations connect. Monthly ROI × 12 = Annual ROI

The baseline is the only thing that makes every number that follows real. Set it before the build begins, not after.

Frequently Asked Questions

How do you calculate automation ROI?
Measure four things before the build: manual hours per week, error rate, response time, and monthly process cost (hours × rate + tools + rework). Then measure the same four things at 30, 60, and 90 days. The before monthly cost minus the after monthly cost is the monthly ROI.
When do you see ROI from AI implementation?
Most teams see measurable results at 30 days, typically 60 to 70 percent reduction in time spent on the automated process. At 60 days the team's ownership changes. At 90 days isolated automations start connecting, and ROI compounds.
What is a realistic ROI from automation?
For desk-level automations, a payback period of one to four months is typical. For team-level rollouts, three to eight months. The size of the ROI depends on the baseline cost of the manual process, which is why measuring before the build matters.
How long does it take to see results from AI implementation?
The first measurable result usually appears in week two or three, when the first workflow goes live and the team sees the time difference. The 30-day review against the baseline produces the first verifiable number.
What metrics should I track for AI implementation?
Manual hours per week, error rate, response time, and monthly process cost. Track all four before the build starts and at 30, 60, and 90 days after go-live. Those four numbers tell the complete ROI story.

Priyankka Wadhwa is the Founder of Let's Execute AI. Her practice works with companies in the United States and India, not as advisors, but as the team that maps, builds, and hands over. She does not deliver strategy. She delivers working systems and the people who can run them.

Start with the baseline.

Two hours. One process. The number you need before anything gets built.

Book the Diagnostic →
· LET'S EXECUTE AI · BOOK A CALL PW