How many engineer hours do your tickets cost?
Every ticket starts at L1. Some escalate to L2, some to L3, and a few pull in the app and infra teams. Each hop adds engineer time. Model your flow below and see what changes when Sherlocks closes tickets itself and stops escalations before they happen.
Your numbers
Fills the With Sherlocks values in the table below. Edit any of them.
Adds net savings, ROI multiple and payback.
Where the hours go, per month
Assumptions by tier
Each cell is today → with Sherlocks. Tickets resolved here is where each ticket's journey ends, so each column should add up to 100%.
| Tier | Tickets resolved here | Engineer time per ticket | Cost per hour |
|---|---|---|---|
Sherlocks (AI) Closes tickets before L1 | 0% 0 → 300 tickets | n/amin human review per ticket it closes | $ reviewer's rate |
L1 support / NOC 1,000 → 700 arrive | % 600 → 450 tickets | min | $ |
L2 operations 400 → 250 arrive | % 250 → 170 tickets | min | $ |
L3 SRE / platform 150 → 80 arrive | % 100 → 60 tickets | min | $ |
App & infra teams 50 → 20 arrive | % 50 → 20 tickets | min | $ |
| Total | 100% → 100% | ||
How the model works
Every hop costs time
A ticket the app team fixes was also triaged at L1, worked at L2 and handed off by L3. The model charges each tier's time to every ticket that passes through it, which is why escalations are the expensive part.
Sherlocks closes tickets first
Sherlocks sits in front of L1. Noise, duplicates and known issues get investigated and resolved there, and a human only reviews the result, so those tickets cost minutes instead of a full triage.
Fewer, faster escalations
The tickets people still work arrive with the root cause already investigated, so each tier closes more of them itself and spends less time on each one.
Why tickets, not incidents?
Every incident raises one or more tickets, but most tickets are not incidents. Alerts that clear on their own, duplicates, and known issues all land in the queue and all take someone's time. Counting only incidents misses most of the work. Counting tickets measures the load your team actually carries.
Run it on your real ticket data
Share a month of ticket exports and we will build this model on your actual escalation paths, then show you which tickets Sherlocks would have closed.
Book a 30-minute walkthrough