How to Calculate Chatbot ROI Before You Commit to a Plan

Chatbot ROI is the monthly cost of answering repetitive questions by hand, minus the plan price, divided by the plan price. That’s it. This guide is for the owner who has to justify a monthly subscription to a partner or manager and has no idea which numbers to put in front of them. At the end you get a fill-in template built from your own inbox, calendar and payroll figures. Not industry averages that nobody at your company can check.

What Goes Into a Chatbot ROI Calculation?

You need three inputs on the benefit side and one on the cost side. Before you look up a single figure, write them down as placeholders:

1. [Q] - repetitive questions your team receives per month
2. [M] - hours spent on a typical answer (a few minutes, written as a fraction of an hour)
3. [H] - fully loaded hourly cost of the person answering
4. [P] - monthly plan price

The fully loaded hourly cost is gross pay plus what the employer pays on top (social contributions, benefits and so on), divided by the hours actually worked. Not the hours on the contract. Every one of these inputs comes from your own records, and that’s the whole point: when someone pokes at a number, you can show where it came from.

How to Measure the Cost of Repetitive Support Questions

Count and time your questions for two to four weeks before you estimate anything. Tag every email, chat and call by question type. Then throw out everything except the questions you answer from content you already have: opening hours, pricing rules, delivery terms, setup steps. Not sure where the line is? Start with repetitive questions that someone answers the same way every single time. And time a sample of those replies, so [M] is measured rather than guessed (people tend to guess low). The monthly cost of repetitive questions is then:

[Q] x [M] x [H]

Putting Plan Prices on the Cost Side

The cost side is the plan price plus the time spent keeping the bot’s content up to date. For [P], take the published monthly plan prices and look at the token allowance each plan includes. Pick the one that fits your question volume and the size of your knowledge base. One thing people miss: voice (spoken questions, spoken answers) is on Professional. Starter is chat-only.

Upkeep counts too. Don’t skip it. Prices change, policies get rewritten, new products show up and need answers. Put it in as [U] hours per month x [H]. If you’re worried that number will balloon, the piece on the upkeep cost of knowledge shows how to keep it small without hiring an editor.

A Worked Chatbot ROI Template You Can Fill In

The template sets benefit against cost: benefit = [Q] x [A] x [M] x [H], cost = [P] + [U] x [H], and ROI = (benefit - cost) / cost. [A] is the share of repetitive questions the bot actually answers. You measure it. You don’t assume it. Go line by line:

1. Enter [Q] from your tagged inbox count.
2. Enter [M] from your timed sample.
3. Enter [H] from payroll.
4. Enter [A] from testing on your own questions.
5. Calculate the benefit: [Q] x [A] x [M] x [H].
6. Calculate the cost: [P] + [U] x [H].
7. Calculate ROI: (benefit - cost) / cost.

Then check break-even, meaning how many answered questions per month it takes for benefit to equal cost: ([P] + [U] x [H]) / ([M] x [H]). Honestly, this is the line I’d lead with in the meeting. A manager can walk through this chatbot cost vs benefit sheet cell by cell, because every cell has a source they can look up themselves.

How to Measure Chatbot Savings After Launch

To measure chatbot savings, swap each estimate for data from the first weeks of real conversations. Before launch, though, run the bot against a list of real questions pulled from your inbox and test before trusting the numbers. That’s your first measured [A]. Rough, but real.

After launch, read the conversation logs regularly. Boring? A bit. But every unanswered question points to a gap in your content, and closing those gaps is how [A] goes up. The guide on what conversation reviews reveal covers what to look for. Last step: compare the repetitive questions in your inbox before and after launch. Same tags as before, so you’re comparing like with like.

Presenting the Chatbot Business Case to a Partner or Manager

One page. Inputs, formula, break-even point, and a proposed review date. A solid chatbot business case covers:

• the data source for each input
• a conservative and a realistic value for [A]
• the estimated upkeep hours
• the date of the first review

Frame the monthly plan as a decision you can reverse after a measured review, not a long commitment. Why does that matter? It lowers the stakes of saying yes, and it gives the skeptic in the room a clear point where they get to decide whether to keep going.

Chatbot ROI is only as convincing as what goes into it. When every figure comes from your own inbox, timesheets and payroll, the chatbot return on investment argument survives questions from anyone reviewing it. Even the careful ones.

FAQ