ITSM AI ROI Calculator
An internal service desk drowns in tickets that follow a runbook, like password resets, access requests and software installs. Runbook work is exactly what AI executes well. Put your team's numbers into the calculator below.
The preset opens with 1,500 IT requests a month, 12 minutes of an engineer's time per request at 35 euros an hour, 900 input and 300 output tokens per resolved request and a person checking 15 percent of the model's work, against a 15,000 euro representative build. Every figure recalculates as you edit.
Your numbers
With taxes and overhead the hour costs €16.25.
Advanced · minutes per task, build, hosting, tokens
Since you opened this page, the task has cost
The same day, two ways
At 1,500 requests a month, that is 72 a working day. A person spends 12 minutes on each.
Each square is one request. The gold sliver is the model writing all 72 answers, about 6 seconds each. The hatched part is the person reading 15 of every 100 before they go out.
Twelve months, the build paid once
The build pays for itself in 32 days. After that every month keeps €4,685. Twelve months, €51,222.
Out of 100 answers, how many does a person read?
Click a square. This is the honest number in the whole calculation, and it costs more than the model does.
Honestly, it barely matters
Seven models at list price, per request with the same checking. Cheapest to dearest, the whole price list fits in €0.0103. The SaaS agent billed per resolution is 9.9× away.
With Claude Sonnet 5.5 the tokens are €0.0053 of the €0.0865 per request. The rest is the person reading. Pick the model for quality, decide the checking for cost.
Your inputs and published prices. Nothing else.
12 min / 60 × €16.25 = €3.25(900 × €1.84 + 300 × €9.20) / 1M × 1.2 = €0.005315% × 2 min / 60 × €16.25 = €0.0813(1,500 × (€3.25 − €0.0865) − €60) × 12 − €5,000 = €51,222All formulas, model prices and task defaults
Every formula with your values
12 min / 60 × €16.25 = €3.25(900 × €1.84 + 300 × €9.20) / 1M × 1.2 = €0.005315% × 2 min / 60 × €16.25 = €0.0813(1,500 × (€3.25 − €0.0865) − €60) × 12 − €5,000 = €51,222€5,000 / €4,685 a month = 32 days(€5,000 / 12 + €60) / (€3.25 − €0.0865) = 151 a monthPublished list prices, USD per million tokens
Prices as of 2026-10-21. Copied from the vendors' pricing pages on that date. Check the linked page before you sign anything.
| Model | Input | Output | Source |
|---|---|---|---|
| Claude Haiku 5.5Anthropic | $0.1 | $0.5 | Pricing page ↗ |
| Claude Sonnet 5.5Anthropic | $2 | $10 | Pricing page ↗ |
| Claude Opus 5.5Anthropic | $4 | $20 | Pricing page ↗ |
| Gemini 3.8 FlashGoogle | $0.75 | $3.75 | Pricing page ↗ |
| Gemini 3.1 ProGoogle | $2 | $12 | Pricing page ↗ |
| GPT-6 LunaOpenAI | $0.1 | $0.5 | Pricing page ↗ |
| GPT-6 SolOpenAI | $2 | $10 | Pricing page ↗ |
| Intercom Fin (per resolution)SaaS, per resolution | $0.99 / task | Pricing page ↗ | |
What each task assumes
| Task | Minutes | Tokens in / out | Checked | Build | Saving a year |
|---|---|---|---|---|---|
| Support ticket reply | 8 | 1,200 / 400 | 20% × 3 min | €4,500 | €90,641 |
| IT request (reset, access) | 12 | 900 / 300 | 15% × 2 min | €5,000 | €51,222 |
| Lead research and qualification | 15 | 2,500 / 600 | 30% × 4 min | €3,500 | €31,543 |
| Sales email or proposal draft | 40 | 3,000 / 1,200 | 100% × 8 min | €3,000 | €27,408 |
| Invoice parsing and entry | 6 | 2,000 / 500 | 25% × 2 min | €6,000 | €46,547 |
| Email classification and routing | 2 | 600 / 100 | 5% × 1 min | €2,500 | €47,247 |
| RFP / security questionnaire answer | 25 | 4,000 / 800 | 60% × 6 min | €5,000 | €8,148 |
| Payment reminder and collections touch | 6 | 1,500 / 300 | 30% × 2 min | €4,000 | €21,486 |
| Your own task | 10 | 1,500 / 500 | 30% × 3 min | €4,000 | €24,749 |
Working estimates from projects delivered in 2025 and 2026, not benchmarks. Every number is editable above. The saving uses your salary, currency and model.
Email me this estimate
One email with the figures and the link to this exact calculation. No newsletter.
What AI actually automates in ITSM
The high-volume, low-judgment tickets are the target. A first ITSM automation typically covers:
- 01Password resets and account unlocks, verified and executed end to end without an engineer.
- 02Access and permission requests routed through approval and provisioned automatically once approved.
- 03Ticket categorization, priority and assignment, so nothing sits in the intake queue overnight.
- 04Answers from your internal docs on VPN setup, device policies and the how-to questions that make up a third of many queues.
- 05Incident summaries and status updates that engineers currently write by hand during and after outages.
Realistic ranges to expect
Published benchmark data puts AI-driven deflection of service desk tickets in the 50 to 70 percent range for the routine categories, The calculator does not apply that rate; it compares the cost of one request done by an engineer with the cost of the same request done by the model plus the checks a person still makes. If your documentation is patchy or your stack is heavily customized, set the review share to 30 or 40 percent and use that as the planning number.
Because IT hours cost more than front-line support, each request the model resolves is worth more, so the same review share pays back faster here than in most departments.
What the research says
Dated, sourced numbers from the two largest ITSM benchmark datasets published recently:
Freshworks' 2025 Freshservice benchmark, built on more than 187 million tickets from 10,551 organizations, found IT teams using its AI copilot saw a 76.6 percent decrease in ticket resolution time, and its AI agent reached a 65.7 percent ticket deflection rate.
In April 2026 ServiceNow published Robinhood's numbers. AI deflects 70 percent of employee requests across IT, HR and Legal before a human needs to step in, saving about 2,200 hours of manual effort across 1,300 tickets a month.
How to read your result
Both benchmark numbers above come from vendors selling the tools, so treat them as the optimistic edge of the range. Our calculator deliberately prices the human review instead of quoting a deflection rate. The honest ITSM story is still strong. The ticket categories are well-defined, the fix procedures are documented and success is trivially measurable in your existing ITSM tool, which makes this one of the safest first AI projects an operations team can run.
ITSM ROI questions
01Does this require replacing our ITSM platform?
No. We build on top of what you run, Jira Service Management, Freshservice, ServiceNow or a shared inbox, through their APIs. The automation reads and writes tickets in the tool your team already uses.
02What about security and access control?
Actions that grant access always go through your existing approval chain; the AI prepares and executes, it does not approve. For regulated environments we deploy on-premise so no ticket data leaves your network.
03Which tickets should we automate first?
Sort last quarter's tickets by category volume and pick the top two runbook categories, usually resets and access. That slice alone often covers the whole build cost within months.
04How do we measure success after launch?
Three numbers from your ITSM tool: deflection rate on the automated categories, median resolution time and reopen rate. We set the baseline before launch so the delta is unarguable.