# Investigation cost calculator: What does production investigation cost your team?

> Calculate one part of the production reliability burden: engineering time lost to triage, repeat investigations, and context switching.
>
> Markdown version of https://cleric.ai/calculator for AI agents and LLMs. Site index: https://cleric.ai/llms.txt

The calculator is interactive: six sliders set your team’s inputs, a results card shows what investigation costs today and what Cleric returns, and three questions unlock a plan for your team. The inputs, formulas, and plan content are below so the calculation can be reproduced.

- Open the calculator: https://cleric.ai/calculator
- Book a demo: https://cleric.ai/demo

## Inputs

| Input | Description | Range | Step | Default | Query param |
| --- | --- | --- | --- | --- | --- |
| Engineers on your team | Total engineers who respond to or are interrupted by investigations. | 2–200 | 1 | 12 | `eng` |
| Avg. engineer fully-loaded cost | Annual salary + benefits + overhead. US average is around $220k–$280k. | $100k–$500k | $10k | $250k | `salary` |
| Investigations per month | Any alert that requires human investigation, P0 through P3. | 5–300 | 1 | 28 | `inc` |
| Avg. engineers pulled per investigation | On-call + secondary responders + interruptions to adjacent teams. | 1–8 | 0.5 | 2.5 | `pulled` |
| Avg. investigation time (hours) | Time from alert to root cause identification, not just resolution. | 0.25–8 | 0.25 | 1.2 | `hrs` |
| % repeat investigations (same root cause) | Investigations you’ve diagnosed before but had to re-run. | 5–80% | 1 | 34% | `repeat` |

The page reads these query parameters to prefill the sliders, for example https://cleric.ai/calculator?eng=12&salary=250000&inc=28&pulled=2.5&hrs=1.2&repeat=34

## Formulas

All figures are annual. The hourly rate assumes 2,080 working hours per year. `repeat` is used as a fraction (34% = 0.34).

**Cost today**

- Hourly rate = salary ÷ 2,080
- Investigation hours = investigations per month × engineers pulled × investigation time × 12
- Repeat hours = investigation hours × repeat %
- Context-switch hours = investigations per month × 12 × engineers pulled × 0.38
- Total hours = investigation hours + context-switch hours
- Total cost = total hours × hourly rate
- FTEs = total hours ÷ 2,080

**What Cleric returns**

Savings are computed at a coverage rate: 50% for the first 90 days and 85% at steady state (6+ months).

- Time-reduction savings = (investigation hours − repeat hours) × 0.75 × coverage × hourly rate
- Repeat-elimination savings = repeat hours × 0.80 × coverage × hourly rate
- Context-switch savings = context-switch hours × 0.60 × coverage × hourly rate
- Annual savings = the sum of the three
- Hours saved = (investigation hours × 0.75 + context-switch hours × 0.60) × coverage
- FTEs recovered = hours saved ÷ 2,080

**Repeat root-cause rate**

- Today = your repeat %
- Month 6 = a range: low end = max(2, round(today × 0.15)), high end = max(4, round(today × 0.30))

Money is shown rounded to the nearest $1k (for example `$160k`), or to one decimal in millions (for example `$1.2M`). FTEs are shown to one decimal.

## Results

The results card has three steps.

**Step 1 · Roadmap lost to investigation:** the total cost “of roadmap work absorbed by investigation” and the FTEs “stuck investigating instead of building”, with a one-line summary of your inputs.

**Step 2 · What Cleric returns to roadmap:** a toggle switches between **First 90 days** and **Steady state** (the default). It shows the annual savings “of roadmap work restored” in that period, the FTEs “back on roadmap, SLAs, and customer revenue”, and the repeat root-cause rate today and by month 6.

**Step 3 · Your plan:** locked until you answer the three questions below (“Reveal your team’s plan.”), then shows the plan for your level.

With the default inputs, the card shows **$160k** and **0.6 FTEs** lost to investigation, **$99k/yr** and **0.4 FTEs** returned at steady state (**$58k/yr** and **0.2 FTEs** in the first 90 days), and “Repeat root-cause rate: 34% today → 5–10% by month 6.”

The page notes: Every trial includes 500 evaluation credits with no time limit. Model based on observed Cleric deployments. Hourly rate = salary ÷ 2,080.

A **Copy share link** button copies a link with your inputs and answers, and a bar fixed to the bottom of the screen summarizes the cost lost to roadmap and what Cleric returns.

## Three questions to unlock your plan.

Each answer scores 1 to 3 points.

1. **When the same root cause recurs, how does your team find out?** Tracked and reviewed (3) · Recognized on-call (2) · Investigated from scratch (1)
2. **After a novel incident, what happens to the investigation context?** Captured in a searchable system (3) · Written up in a postmortem (2) · Stays with whoever investigated (1)
3. **What share of senior engineers’ time goes to investigation vs. architecture?** Mostly architecture (3) · Split evenly (2) · Mostly firefighting (1)

The total (3 to 9) sets your level: 3–4 is Level 1, 5–7 is Level 2, and 8–9 is Level 3.

### Level 1 of 3 · Alert noise: Your team is fighting alert noise.

Most alerts are noise. Cleric helps your team find the signal before automating anything.

How Cleric helps your team:

1. **Audits your alert rules.** Flags noise, duplicates, and orphaned rules in your existing tools. Nothing goes without your review.
2. **Maps tool overlap.** See what overlaps. Consolidate with confidence, not guesswork.
3. **Holds off on autonomy.** Stays in audit mode until signal is clean. No babysitting false resolutions.

- See what Cleric surfaces in your stack: https://cleric.ai/demo
- Read: What’s good for AI agents is good for engineers: https://cleric.ai/blog/whats-good-for-ai-agents-is-good-for-engineers

### Level 2 of 3 · Alert fatigue: Your team is mostly firefighting.

Investigation volume is crowding out everything else. Cleric runs first-pass on every alert so your engineers only see what needs them.

What Cleric takes off your plate:

1. **Investigates every alert first.** Analyzes logs, metrics, traces, and recent deploys. Your engineers see a root cause, not a blank page.
2. **Closes known patterns.** Recurring investigations resolve without paging anyone. Novel incidents are where your team spends time.
3. **Drops engineers-per-incident to one.** The brief includes root cause and confidence. One reviewer decides remediation; no war rooms.

At your volume, Cleric absorbs ~N investigations/month, giving your team back ~H of focus time. Here N = investigations per month × 0.85 (rounded) and H = N × investigation time × 0.75 hours (rounded). With the default inputs: ~24 investigations/month and ~22 hrs.

- See Cleric investigate a real incident: https://cleric.ai/demo
- Read: Why your engineers are drowning in alerts: https://cleric.ai/blog/why-your-engineers-are-drowning-in-alerts

### Level 3 of 3 · Ready to scale: Your team is past firefighting.

Low load, clean signal. Cleric scales your team’s judgment across production.

What Cleric does alongside you:

1. **Pairs with your coding agents.** They ship faster. Cleric catches breaks so your team stays in flow.
2. **Executes the SOPs your team defines.** Your engineers set the standards; Cleric runs them. Review traces, tune policy.
3. **Reports what actually matters.** Repeat-RCA rate, novel vs. known, time-to-auto-resolve — not MTTR theater.

- Book a consultation: https://cleric.ai/demo
- Read: Agents will be ICs, we will be their EMs.: https://cleric.ai/blog/agent-engineering-is-not-software-engineering