AI chatbot ROI calculator
Estimate what a customer-support chatbot is worth to your team. Set your ticket volume, handle time, and support cost, pick a deflection rate, and see tickets deflected, hours saved, net savings, and ROI update live.
Tickets deflected / mo
200
Support hours saved / mo
26.7
Gross savings / mo
$746.67
Net savings / mo
$506.67
ROI multiple
2.1×
Monthly support cost, before vs. after
Deflection of 30–50% is a typical starting range for chatbots grounded on real documentation, and it improves with tuning. Start with conservative inputs — an estimate you can beat is worth more than one you have to defend.
How support-chatbot ROI is calculated
The model works up from the tickets a bot handles alone. First, deflected tickets = ticket volume × deflection rate. Each deflected ticket saves your fully-loaded cost per ticket, which itself is roughly average handle time × loaded hourly cost. So monthly savings = deflected tickets × cost per ticket − chatbot cost, and ROI = (savings − chatbot cost) ÷ chatbot cost. The two levers that dominate the result are ticket volume and deflection rate — everything else scales those two.
What each input means
- Ticket volume. How many support contacts you receive in the period. It is the base every other number multiplies against, so a bot on a high-volume queue pays back far faster than the same bot on a quiet one.
- Average handle time. The minutes an agent spends per contact end to end. Combined with a loaded hourly cost, it produces your cost per ticket — the amount each deflection actually saves.
- Deflection rate. The share of tickets the bot fully resolves without a human. This counts genuine resolutions only, not conversations where the user gave up. It is the single most sensitive input in the whole model.
- Cost per contact. Your fully-loaded cost to handle one ticket — labor plus benefits, tooling, and management overhead. Use a figure your finance team recognizes so the case survives scrutiny.
What realistically moves deflection
Deflection climbs when the bot is grounded on your real documentation through retrieval (RAG) rather than answering from generic knowledge, when it covers the handful of repetitive questions that make up the bulk of your queue, and when it has clean escalation paths — a bot that hands hard questions to a human quickly earns the trust to keep the easy ones. Be honest about the caveats, though. Deflection varies widely by use case, so measure it on your own traffic instead of assuming a benchmark. A deflected ticket is not automatically a resolved one: if users abandon and churn rather than escalate, the "savings" can cost you revenue, so track CSAT alongside deflection. And a poorly built bot that frustrates people can add cost and harm satisfaction — the technology is not free ROI, only a lever you have to earn.
Beyond the cost savings
The hard-dollar saving is the number to build a business case on, but a good support bot also delivers benefits that do not show up directly in the ROI figure: instant responses at any hour, round-the-clock coverage without extra headcount, and agents freed from repetitive tickets to focus on the complex, high-value cases only a human can handle. Treat these as upside layered on top of the savings — real, worth mentioning, but secondary to the concrete number the calculator produces.
Frequently asked questions
- How do you calculate chatbot ROI?
- Start from the tickets the bot resolves without a human: deflected tickets = ticket volume × deflection rate. Each of those saves your fully-loaded cost per ticket, so monthly savings = deflected tickets × cost per ticket. ROI compares that saving to what the bot costs to run: ROI = (savings − chatbot cost) ÷ chatbot cost, usually expressed as a percentage or a payback period. The calculator above does this live as you adjust the inputs.
- What is deflection rate?
- Deflection rate is the share of incoming contacts the chatbot fully resolves on its own, so a human agent never has to touch them. A 30% deflection rate on 10,000 monthly tickets means 3,000 were handled end-to-end by the bot. It is the single biggest lever in the model — but it only counts genuine resolutions, not conversations where the user gave up or silently escalated elsewhere.
- What is a realistic support-bot deflection rate?
- It varies enormously by use case, so treat any single number with suspicion. Deflection tends to be higher when your ticket mix is dominated by a handful of repetitive, well-documented questions (password resets, order status, policy lookups) and lower when tickets are novel, account-specific, or require an action the bot cannot take. The honest answer is to pilot on your real traffic and measure, rather than plugging in a benchmark you read somewhere.
- How much does an AI chatbot save per ticket?
- The saving per deflected ticket is your fully-loaded cost per ticket — the labor and overhead a human would have spent handling it. That figure differs widely between a low-cost, high-volume support team and a specialized one with long handle times, so the calculator asks you to supply your own cost per ticket (or derive it from handle time and hourly cost) instead of assuming a figure for you.
- How do I estimate cost per ticket?
- A common rule of thumb is cost per ticket ≈ average handle time × fully-loaded hourly cost. Average handle time is the minutes an agent spends per contact end to end; fully-loaded hourly cost should include salary plus benefits, tooling, and management overhead, not just base pay. If your finance team already tracks a cost-per-contact number, use theirs — it will hold up better in a business case.
- Does a chatbot do anything beyond saving money?
- Yes, though those benefits are softer than the hard-dollar number. A bot answers instantly at any hour, so response and resolution times drop and coverage extends to nights and weekends without extra headcount. It also frees agents from repetitive tickets to focus on complex, high-value cases. Treat these as upside on top of the savings figure, not a substitute for it.
- Is this calculator free?
- Yes — it runs entirely in your browser, needs no signup, and sends nothing to a server. Every input is editable, so you can model your own ticket volume, handle time, deflection rate, and cost per ticket, and stress-test the numbers before taking them to anyone.
Related tools: LLM API cost calculator (price the model behind the bot), all free tools, and the Shopify integration for storefront support.
More free tools
- LLM price comparison — Live table of every model's API prices — sort, filter, and cost out your own request size.
- LLM cost calculator — Estimate per-request, daily, and monthly spend from token counts and request volume.
- RAG cost calculator — Full pipeline costs — embeddings, vector storage, and generation — from corpus size and query volume.
- RAG chunking visualizer — Paste a document and watch chunk size, overlap, and strategy change the chunks before you embed.
- LLM context window comparison — Count the tokens in any text and check which models it fits, against live context windows.
- Tool call & MCP schema builder — Build function-calling parameters once, copy valid JSON Schema for OpenAI, Anthropic, and MCP.