docs(TSP-1320): add workforce evaluation documentation

[claude]

Summary

• Updated build/agents/build-your-agent/evals.mdx to clarify that the Evaluate tab works for both individual Agents and Workforces
• Added a Node field row to the Tool Usage check type table describing the node picker for Workforce evals
• Expanded step 7 (Tool simulations) to explain per-node simulation behavior when running Workforce scenarios, including node disambiguation for duplicate Agent nodes
• Added a new "Evaluating Workforces" section covering the Evaluate tab on the Workforce canvas, Tool Usage check node scoping, and per-node tool simulation
• Updated build/workforces/create-a-workforce.mdx to add a cross-reference link to the Evals page in the "What's next?" section

Linear issue

https://linear.app/relevance/issue/TSP-1320/

[linear]

TSP-1320

[mintlify]

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

Project	Status	Preview	Updated (UTC)
relevanceai	🟢 Ready	View Preview	Jun 24, 2026, 2:27 AM

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

[github-actions]

🎯 Vibe check

Reviewed: 2 files (2 with issues)

Scores

	Dimension	Score	What's holding it back
🟡	Consistency	8/10	Bold label (**Rollout Status**:) inside a callout in evals.mdx violates CLAUDE.md's "no bold labels inside callouts" rule; two headings use lowercase product terms (checks, tool) where the rest of the file capitalizes them.
🟢	Technical clarity	9/10	No issues found. Field tables are thorough, UI references are exact, and edge cases (truncation, duplicate Workforce nodes, per-node simulation) are well documented.
🟢	Non-technical clarity	9/10	No issues found. Both pages define terms before using them and give concrete examples.
🟡	Structure	8/10	create-a-workforce.mdx uses a plain numbered list for an 8-step sequential procedure; CLAUDE.md calls for <Steps> on sequential procedures.

Score key: 🟢 9–10, 🟡 6–8, 🔴 1–5.

✨ Overall vibe: The evals page is the stronger of the two — comprehensive, logically layered, and precise about edge cases like conversation truncation and duplicate Workforce nodes. Both files are close to ship-ready; the three fixes below are quick wins that bring them into full compliance with the content standards.

🔧 Issues (3)

• build/agents/build-your-agent/evals.mdx:8 — **Rollout Status**: is a bold label acting as a mini-heading inside an <Info> callout. CLAUDE.md explicitly bans bold labels inside callouts. Drop the label and write the paragraph directly: "Evals is rolling out progressively, starting with Enterprise customers. If you don't see this feature in your account yet, reach out to your account manager to discuss access."

• build/agents/build-your-agent/evals.mdx:215 — Heading ### Tool Usage checks with node scoping uses lowercase checks. Every other reference to this product feature in the file uses Checks (capital C). Should be: ### Tool Usage Checks with node scoping.

• build/agents/build-your-agent/evals.mdx:219 — Heading ### Per-node tool simulation uses lowercase tool. The body text on lines 147, 149, 152, and 221 all use Tool (capital T) when referring to the product feature. Should be: ### Per-node Tool simulation.

🧩 Component suggestions (1)

• build/workforces/create-a-workforce.mdx:20–28 — The 8-step create procedure is a plain numbered list. Per CLAUDE.md, sequential procedures should use <Steps> for the visual progress indicators. Wrap steps 1–8 in a <Steps> component:

  <Steps>
    <Step title="Open Workforce">Click **Workforce** in the left sidebar.</Step>
    <Step title="Create canvas">Click **+ New Workforce** to open the visual canvas.</Step>
    ...
  </Steps>

  The image on line 29 can stay outside <Steps> after the last step.

✅ Clean files (0)

No files were fully clean, though build/workforces/create-a-workforce.mdx has only one component suggestion and is otherwise solid.

🔋 Credit usage

Item	Count
Files reviewed	2
Context pages read	2
Total lines processed	~701

Files read: build/agents/build-your-agent/evals.mdx (407 lines), build/workforces/create-a-workforce.mdx (46 lines), get-started/core-concepts/workforces.mdx (126 lines), build/agents/build-your-agent/triggers.mdx (122 lines)