⚠️

AI Agent Outage Cost Calculator

Estimate revenue loss, SLA penalties, support surge cost, and churn impact when an AI agent goes offline.

Scenario presets

Outage assumptions

Revenue model Monthly revenue / MRR ($) Monthly successful agent outputs Average value per output ($) Outage duration (hours) Incident type SLA penalty rate (% of monthly fee) Churn risk from outage (%) Fallback human hourly cost ($) Human handle time per disrupted request (min) Engineering recovery hours Engineering hourly cost ($)

Estimates are directional. Last updated: 2026-07-07. See notes.

Revenue loss

—

SLA penalties

—

Support surge cost

—

Total outage cost

—

Churn impact

—

Recovery engineering cost

—

Cost per minute of outage

—

Incident-type multipliers

Incident type	Revenue multiplier	Churn multiplier	SLA multiplier	Typical cause
API provider downtime	1x	1x	1x	Frontier LLM API outage or rate-limiting
Model degradation / errors	0.6x	1.3x	0.5x	High error rate, bad outputs, or refusal loops
Deployment / rollback failure	1.2x	1.4x	1x	Bad release breaks agent workflows
Infrastructure crash	1x	1.2x	1x	Server, container, or database failure
Critical tool integration outage	0.8x	1.1x	0.5x	Search, payments, CRM, or data source down
Planned maintenance	0.4x	0.3x	0x	Scheduled updates with notice

Verdict

—

Frequently asked questions

What counts as an AI agent outage?

Any period when the agent cannot accept or complete requests: API downtime, model errors, rate-limiting, deployment failures, infrastructure crashes, or degradations that prevent normal operation.

How is revenue loss estimated?

We convert your revenue model into a per-minute value. For transaction agents, we use conversions per minute × value per conversion. For subscription agents, we use MRR per minute × churn risk. Support agents use cost per disrupted request.

What are SLA penalties?

SLA penalties are contractual commitments to customers when availability falls below a promised threshold. They are often a percentage of fees but rarely cover lost business.

How is support surge cost calculated?

When the agent is down, requests that would have been automated must be handled by humans or queued. We estimate that surge using your fallback handle time and labor cost.

Does this include customer churn?

Yes — you can set a churn risk percentage. We apply it only to revenue exposed during the outage window, not your whole customer base.

What is the difference between planned and unplanned downtime?

Planned downtime is scheduled maintenance and usually has lower churn/penalty impact. Unplanned downtime hurts trust more and often triggers SLA penalties and emergency engineering time.

SLA / Uptime Calc → Human Escalation Cost Calc → Startup Runway Calc → Agent Stack Builder →

Outage cost estimates are directional. They assume linear revenue exposure during downtime and do not include brand damage, legal exposure, or long-tail churn beyond the selected churn risk percentage.