AI SRE comparison
WHAWIT vs Cleric
We admire the transparency. We disagree that the AI should clock out after the diagnosis.
Short answer
Cleric is the category's transparency leader — published pricing, published investigation volumes, and a verification engine that checks findings against the live environment before presenting them. It is also read-only by design: fix proposals arrive as PRs, humans execute changes, and there is no incident platform, no on-call and no pager underneath. WHAWIT shares the conviction that trust is earned with receipts — every claim in its investigations cites its evidence, and every model call is itemized in a ledger — but draws the line differently on action: agents hold rostered shifts, recovery actions execute behind approval gates you configure, and the coding agent's PR arrives implemented and tested, not proposed. Transparency about the diagnosis is good. Transparency about the whole loop is better.
Side by side
| Capability | WHAWIT | Cleric |
|---|---|---|
| AI root-cause investigation with cited evidence | ||
| Reads your observability stack read-only, no migration | ||
| Incident management system of record: lifecycle, timeline, postmortems | ||
| Native on-call: schedules, rotations, escalation policies | ||
| Paging over voice call, WhatsApp, SMS, Slack, Teams, Discord and email | ||
| AI agents with scheduled shifts on the on-call roster | ||
| Coding agent ships the fix as a pull request | ||
| Approval-gated infrastructure remediation | ||
| Itemized AI spend, per analysis, visible in the product | ||
| Published list pricing |
Based on each vendor's public documentation as of August 2026. Cleric is a trademark of its respective owner and is not affiliated with WHAWIT.
Where the products actually differ
Read-only by design vs. guardrailed by design
Cleric's answer to trust is to never act. WHAWIT's is to act inside guardrails you configure: approvals, severity thresholds, cooldowns and daily caps on recovery actions, and human review on every PR. Trust through control beats trust through abstinence — the incident still needs fixing.
A platform under the agent
Cleric assumes an incident stack exists around it. WHAWIT ships it: lifecycle, timeline, postmortems, on-call schedules, escalation policies and seven paging channels, with the investigation attached to the page.
Matched on transparency, honestly
Cleric publishes list pricing; we publish our model instead — AI consumption at cost, itemized per analysis in the product, with the fee covering seats, support and licensing. Their number is on a page; ours is in your billing screen, per analysis, as it happens.
Tested fixes, not proposed ones
WHAWIT's coding agent clones the repo, implements the fix, runs the tests in a sandbox and opens a review-ready PR linked to the incident. A proposal you still have to build is a head start; a tested PR is a finish line.
Questions teams ask
Is WHAWIT an alternative to Cleric?
Yes, with a structural difference: Cleric is an investigation agent that layers on top of an incident platform you still have to run, while WHAWIT is the platform and the agent in one — incident lifecycle, native on-call and multi-channel paging, AI investigation with cited evidence, approval-gated remediation, and a coding agent that opens the tested fix as a pull request.
Can I evaluate WHAWIT against Cleric side by side?
Yes, and you should. WHAWIT connects read-only to the observability stack you already run, so both products can investigate the same incidents during a pilot. We publish the bake-off rubric we ask prospects to score us on at whawit.ai/ai-sre — seven scenarios, ten dimensions, scored blind on your own outages.
Does WHAWIT execute remediation, or only suggest fixes?
Both, with a human in control. Operational fixes run as recovery actions — signed calls to endpoints and MCP tools you connect, gated by approvals, severity thresholds, cooldowns and daily caps. Code fixes arrive as pull requests that a human reviews and merges.
Run them side by side
Keep Cleric paging while WHAWIT investigates the same incidents. Compare the two on your own outages before you move a single rotation.