Update docs: Invent proactively offers alarm auto-diagnosis

[claude]

Summary

Documents the changes from PR #17497 where Invent now proactively offers and explains the Auto-diagnosis by Invent capability during alarm configuration.

Three updates across two files:

• build/invent/invent.mdx — Configure an alarm tab: Expanded the "Turn on auto-diagnosis" section to describe that Invent proactively offers this option, explains what it does, shows a SETUP REQUIRED warning when no alert channel is configured, and guides users to set one up first. Restructured the tab so the alarm explanation comes before the feature description.

• build/invent/invent.mdx — Alerts accordion: Added a note that an active email or Slack alert channel is also required to receive auto-diagnosis findings when alarms fire (not just for scheduled check-ins).

• build/agents/build-your-agent/evals.mdx — Monitor section: Added a callout explaining that alarms with Auto-diagnosis by Invent can be configured from Monitor dashboards, with a link to the full documentation on the Invent page.

Relates to: https://linear.app/relevance/issue/TSP-1429/

[mintlify]

Preview deployment for your docs. Learn more about Mintlify Previews.

Project	Status	Preview	Updated (UTC)
relevanceai	🟢 Ready	View Preview	Sep 1, 2026, 4:36 AM

💡 Tip: Enable Workflows to automatically generate PRs for you.

[linear]

TSP-1429

[github-actions]

🎯 Vibe check

Reviewed: 2 files (2 with issues, 0 clean)

Scores

	Dimension	Score	What's holding it back
🟡	Consistency	6/10	evals.mdx: "agents"/"agent" appears lowercase 7+ times in the "Why use Evals?" section when referring to the Relevance AI product — should be "Agents"/"Agent". Bold label inside <Info> callout violates CLAUDE.md. Subject-verb error ("Evals enables"). Terminology mismatch across files: evals.mdx calls the monitoring feature "Monitor dashboards" while invent.mdx calls the same UI element "Performance dashboards".
🟡	Technical clarity	8/10	invent.mdx: two anchor links (#alerts) point to an <Accordion title="Alerts">, not a heading — accordion titles don't create page anchors in Mintlify, so these links won't resolve.
🟡	Non-technical clarity	8/10	evals.mdx: "Why use Evals?" section switches to first-person plural ("we are using agents…we want to be sure") and reads more like a pitch than documentation. Informality makes it hard to skim.
🟢	Structure	9/10	Both pages are well-organised and logically layered. evals.mdx is a comprehensive concept+how-to page that ends at the FAQ with no closing CTA — minor for a page this long, but worth adding.

Score key: 🟢 9–10, 🟡 6–8, 🔴 1–5.

✨ Overall vibe: Both pages are thorough and well-structured; invent.mdx is notably clean. The main work is in evals.mdx: a "Why use Evals?" section that reads like a sales paragraph (informal voice, uncapitalized product names, a grammar error) undercuts an otherwise precise page. Aligning the dashboard terminology between the two files ("Monitor dashboards" vs "Performance dashboards") is the highest-priority fix since readers will move between these pages.

🔧 Issues (5)

• build/agents/build-your-agent/evals.mdx:8 — <Info> callout contains **Rollout Status**: as a bold label prefix. CLAUDE.md prohibits bold labels inside callouts — it must be a single plain paragraph. Rewrite as: Evals is rolling out progressively, starting with Enterprise customers. If you don't see this feature yet, reach out to your account manager to discuss access.

• build/agents/build-your-agent/evals.mdx:55–58 — "agents" and "agent" appear lowercase 7 times in this section while referring to the Relevance AI product feature. All should be "Agents"/"Agent". Affected phrases: "we are using agents", "your agents are hitting the mark", "testing agents across", "agent cost effectiveness", "deploy agents", "agents can run effectively", "have agents focus".

• build/agents/build-your-agent/evals.mdx:57 — Subject-verb agreement: "Evals enables testing agents" → "Evals enable testing Agents". ("Evals" is plural.)

• build/agents/build-your-agent/evals.mdx:55 — First-person plural ("we are using agents we want to be sure") is out of place in technical documentation. Rewrite in second person: "When you use Agents, you want confidence that they are hitting your business outcomes."

• build/invent/invent.mdx:195, 300 — Both [Alerts](/build/invent/invent#alerts) and [Invent alert channel](/build/invent/invent#alerts) target an <Accordion title="Alerts">, not a heading. Accordion titles don't create HTML anchors in Mintlify, so these links silently fail. Promote the Alerts accordion to a proper ### Alerts heading (removing the accordion wrapper if the content is short), or update both links to target the nearest real heading above it (#manage-performance or similar).

⚠️ Contradictions (1)

• build/agents/build-your-agent/evals.mdx:300, 302, 316, 334 calls the feature "Monitor dashboards" throughout the Monitor section. build/invent/invent.mdx:30, 262, 293 calls the same UI concept "Performance dashboards" (e.g., "Alarms evaluate the scores collected by a Performance dashboard"). Both pages are describing the dashboards created in Evaluate → Monitor on an Agent — they should use the same term. evals.mdx owns the Monitor section, so its name ("Monitor dashboards") is the more precise one; invent.mdx should align.

🏗️ Page structure (1)

• build/agents/build-your-agent/evals.mdx — Page ends at the FAQ with no closing CTA. This is a substantial concept+how-to page, and a reader who's just learned about Evals needs somewhere to go. Suggest adding a brief ## What's next? section pointing to /build/invent/invent ("Invent can build your initial Evals suite and investigate failing Checks") and /build/agents/build-your-agent/alerts ("Set up alerts to catch issues in live Agent tasks").

🔋 Credit usage

Item	Count
Files reviewed	2
Context pages read	2
Total lines processed	~1,013

Files read: build/agents/build-your-agent/evals.mdx (451 lines), build/invent/invent.mdx (362 lines), build/agents/build-your-agent/version-history.mdx (89 lines), build/agents/build-your-agent/alerts.mdx (111 lines)