Most pitches for AI and automation promise big returns and show no math. You can estimate the return yourself, before spending anything, with numbers you already know or can measure in a couple of weeks.
The basic formula
ROI (return on investment) is the gain minus the cost, divided by the cost. Payback time is the cost divided by the monthly gain. The second one is usually the more useful for a small business, because it tells you how long until the project has paid for itself.
Illustrative numbers only, to show the math. They are not RushRush results or a promise. Use your own numbers.
- Project cost: $3,000 (one time).
- Monthly gain: 6 hours a week saved, at $25 an hour of real cost to you, is about $650 a month.
- Payback: $3,000 ÷ $650 is about 4.6 months.
- ROI after 12 months: ($7,800 − $3,000) ÷ $3,000 = 160%.
Where the gain comes from
There are four places to look. Time saved, counted as hours multiplied by what an hour really costs you (pay plus taxes and benefits, or the value of your own time). Revenue gained, such as leads answered faster or customers who come back more often. Mistakes avoided, like double bookings or wrong invoices. And cash that arrives sooner, for example when contracts get signed in a day instead of a week.
Count every cost
The build is only part of it. Add monthly subscriptions, usage fees, the time you or your team spend reviewing the results, training, and maintenance when a tool changes. Many projects look great until the ongoing costs are added.
What research suggests, and its limits
A study of about 5,000 customer support agents at one software company found that an AI assistant raised the number of issues resolved per hour by 13.8% on average, and by 35% for the least experienced workers. That is one company, one tool and one type of work, so it shows that gains are possible, not what you will get. McKinsey Global Institute estimated in 2017 that about 60% of occupations have at least 30% of their activities that could be automated with existing technology. That tells you where to look, not how much you will save. And a 2021 Zapier survey of 2,000 US small and mid-size business workers found that nearly 70% said automation made them more productive, but that is self-reported opinion.
Measure before you build
For two to four weeks, write down the baseline: hours spent on the task, how long customers wait for an answer, how many leads come in and how many become clients. Without the "before", you cannot prove the "after", and you cannot tell whether the project worked.
When it does not pay
Automation rarely pays for tasks that happen a few times a year, for processes that change every week, or when the monthly cost is close to the value of the time saved. Saying no to those projects is part of a good return. If you are unsure where to start, the AI readiness audit exists for exactly that.