Bot Estate Audit
Your bots cost more than their licenses. Find out how much.
A free, 2-week audit of your automation estate: a fragility score for every bot, a run-cost model, and the processes computer-use agents should take over first.
- 85.2% on OSWorld with Holo4 27B, at $0.08 per task
- SOC 2 Type II
- Read-only by default, outbound-only connection
A fact base before your next renewal
Most automation teams know their license cost. Few know what each bot costs to keep alive. The audit puts a number on both.
01
Fragility score per bot
How often each bot breaks, why, and how many hours your team spends fixing it.
02
Run-cost and savings model
License, infrastructure and maintenance cost per bot, next to what an agent would cost to run the same process.
03
Top 10 bots to retire first
The processes where an agent that reads the screen like a person would stop the break-fix cycle.
04
A readout with CFO-ready numbers
One session with your automation lead and finance partner, with the model and assumptions you can reuse.
Two workshops. A readout in week 2.
Day 1
Share your bot inventory
An export from your orchestrator: bot names, schedules, run history and failure logs. No access to production systems.
Week 1
Two 60-minute workshops
We walk through your highest-cost processes with the people who maintain them, then size the agent alternative.
Week 2
Readout and recommendation
Fragility scores, cost model and a shortlist of bots to retire, with a proposal for a paid prototype on the top two or three.
A moved button shouldn't be an outage
Rules-based bots follow coordinates and selectors. When a screen changes, they stop. Computer-use agents look at the screen, find what they need, and carry on.
| Situation | Scripted bots | H computer-use agents |
|---|---|---|
| Screen or layout changes | Bot fails, ticket raised | Agent adapts and continues |
| Citrix, mainframe, desktop apps | Brittle or out of scope | Works on any UI, no API needed |
| Unstructured inputs | Routed to a person | Read and handled in the flow |
| Judgment calls | Not possible | Routed to a reviewer at set thresholds |
The audit is built for teams that
- 01
Run 50 or more production bots
Large enough that maintenance shows up as a real line item.
- 02
Spend more time fixing bots than building them
Center of excellence headcount up, automation coverage flat.
- 03
Have a license renewal in the next two quarters
The readout gives you a fact base before you sign.
Book a 20-minute scoping call
We confirm the size of your bot estate, agree the success criteria and schedule the two workshops. The audit is free and there's no commitment beyond it.
Useful to have on the call
Approximate number of production bots
Your next license renewal month
The two or three processes that break most often
What teams ask before they start
Yes. We run it with one solutions engineer and one forward deployed engineer for about five days. If the numbers make sense, the next step is a paid prototype on two or three bots. If they don't, you keep the model.