docs(TSP-1419): add documentation for Slack wizard button type

[claude]

Summary

Adds a new "Agent-sent Slack cards" section to the Slack integration page (integrations/popular-integrations/slack.mdx), positioned between the Triggers section and the Escalations section.

The new section covers:

• What agent-sent Slack cards are and when agents send them
• The three button types: message, form, and wizard
• Wizard button type details: multi-step modal navigation, automatic draft persistence, optional typed confirmation step, and single-message delivery of all collected answers
• Early access callout matching the existing "Agent Notifications" beta format
• Technical details accordion for workspace admins (config flag, concurrency controls, cross-region relay)

Relates to: https://linear.app/relevance/issue/TSP-1419/

Test plan

• Verify the new section renders correctly in Mintlify preview
• Confirm <Warning> callout matches the "Agent Notifications" section format
• Confirm <Accordion> for technical details renders and collapses correctly
• Check all internal links (/get-started/support) resolve correctly
• Verify sentence case on all headings
• Confirm section ordering: Triggers → Agent-sent Slack cards → Escalations

🤖 Generated with Claude Code

[mintlify]

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

Project	Status	Preview	Updated (UTC)
relevanceai	🟢 Ready	View Preview	Aug 27, 2026, 7:27 AM

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

[linear]

TSP-1419

[github-actions]

🎯 Vibe check

Reviewed: 1 file (1 with issues, 0 clean)

Scores

	Dimension	Score	What's holding it back
🟡	Consistency	6/10	Five heading-case violations in the Advanced Settings subsection; four instances of "agent" when the product term "Agent" is required; "seamless" in the description (banned-word family); "team are" (British collective-noun agreement); "setup" used as a verb twice.
🟡	Technical clarity	8/10	The page documents the same feature (## Agent-sent Slack cards) twice with overlapping but non-identical content — a reader will be confused about which section is authoritative. <Tip> callout contains a numbered list, which breaks the component contract.
🟡	Non-technical clarity	8/10	Introductory definitions are clear. The duplicate section hurts non-technical readers most — they'll hit the wizard explanation at line 158 and then hit a different wizard explanation at line 255 without knowing one supersedes the other.
🟡	Structure	6/10	The duplicate ## Agent-sent Slack cards heading (lines 143 and 239) is a serious structural flaw — effectively two competing drafts living in the same page. <Warning> used for a routine beta notice (line 197) instead of <Note>. Accordion content contains a numbered list where prose is required.

Score key: 🟢 9–10, 🟡 6–8, 🔴 1–5.

✨ Overall vibe: The page covers real, useful content and the new wizard button type documentation is detailed and well thought-through — but it was added without consolidating the earlier ## Agent-sent Slack cards section, leaving the page with a split personality. Fix the duplicate section first; the heading-case and capitalization cleanup is mechanical but thorough.

🔧 Issues (15)

• integrations/popular-integrations/slack.mdx:3 — "seamless communication" in the description — "seamless" is the adjective form of the banned word "seamlessly". Replace with plain description of what the integration does: e.g. "connect your Agents with Slack so they can respond to messages, post notifications, and trigger workflows".

• integrations/popular-integrations/slack.mdx:8 — "The Relevance AI team are constantly" → "The Relevance AI team is constantly". American English treats collective nouns as singular.

• integrations/popular-integrations/slack.mdx:77–82 — <Tip> contains a numbered list with bold labels. CLAUDE.md rule: callouts must be one short paragraph — no bullet lists, no multi-line content, no bold labels. Convert to a plain sentence or two, e.g. "For channels, run /invite @Relevance AI in the channel. For your DM, just send the Relevance AI bot any message to establish the connection." Drop the Tip wrapper entirely if the content doesn't fit a paragraph.

• integrations/popular-integrations/slack.mdx:101 — "Once setup, you can trigger" → "Once set up, you can trigger". "Setup" is a noun; "set up" is the verb/past-participle form.

• integrations/popular-integrations/slack.mdx:105 — heading ### Advanced Trigger Settings → ### Advanced trigger settings. "Settings" is not a proper noun. "Trigger" follows the sibling page (microsoft-teams.mdx) convention of lowercase in headings.

• integrations/popular-integrations/slack.mdx:107 — heading #### Live Status Updates → #### Live status updates.

• integrations/popular-integrations/slack.mdx:111 — heading #### Exclude Keywords → #### Exclude keywords.

• integrations/popular-integrations/slack.mdx:123 — heading #### No Agent Reply → #### No Agent reply. "Agent" stays capped (product term); "reply" does not.

• integrations/popular-integrations/slack.mdx:136 — heading ### Customize Message Formatting → ### Customize message formatting.

• integrations/popular-integrations/slack.mdx:144 — "when an agent responds to a user" → "when an Agent responds to a user". This refers to the Relevance AI product feature.

• integrations/popular-integrations/slack.mdx:155 — "building an agent that uses" → "building an Agent that uses".

• integrations/popular-integrations/slack.mdx:183 — "To escalate your agent to Slack" → "To escalate your Agent to Slack".

• integrations/popular-integrations/slack.mdx:197–199 — <Warning> used for "This feature is currently in beta for some users". A beta notice is informational, not a risk or irreversible-action warning. Use <Note> instead.

• integrations/popular-integrations/slack.mdx:241 — "When an agent sends a message to Slack" → "When an Agent sends a message to Slack".

• integrations/popular-integrations/slack.mdx:331 — accordion title "When I setup a Slack trigger, I can't find channels in my workspace." → "When I set up a Slack trigger, I can't find channels in my workspace.".

🧩 Component suggestions (1)

• integrations/popular-integrations/slack.mdx:309–316 — the accordion "Can I use my own Slack DM as a trigger?" opens with "Yes! You can use..." and then embeds a numbered list with a bold **Note:** label. CLAUDE.md says accordion content should use flowing sentences, not bullet lists. Rewrite as prose: "Yes. Open a DM with the Relevance AI bot and send it any message — this establishes the connection. Your DM will then appear in the channel selection dropdown in Relevance AI. The /invite command is not required for DMs." The bold Note inside can become a plain sentence in the flow.

🏗️ Page structure (1)

• integrations/popular-integrations/slack.mdx:143 and integrations/popular-integrations/slack.mdx:239 — The heading ## Agent-sent Slack cards appears twice. The first instance (lines 143–175) documents button types in a bullet list and then dives into wizard-specific detail, early-access warning, and a technical accordion. The second instance (lines 239–275) starts over with a <CardGroup> overview of all three button types, then adds wizard usage guidance and an accordion example. These are two competing drafts of the same feature documentation. Merge them into one ## Agent-sent Slack cards section: lead with the <CardGroup> (from line 243, it's the better entry point), follow with a ### Wizard button type subsection that consolidates the detail from lines 158–174, keep the early-access <Warning> and the technical <Accordion>, then add the usage example accordion. Remove the duplicate heading entirely.

⚠️ Contradictions (0)

No contradictions found between the changed file and the context pages read.

🔋 Credit usage

Item	Count
Files reviewed	1
Context pages read	2
Total lines processed	~976

Files read: integrations/popular-integrations/slack.mdx (392 lines), integrations/popular-integrations/microsoft-teams.mdx (504 lines), build/workforces/build-an-ai-workforce/add-triggers.mdx (80 lines, partial)