Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.heymilo.ai/llms.txt

Use this file to discover all available pages before exploring further.

AI → Interview Evaluation is where you set workspace-wide language proficiency rubrics: how Milo scores candidates’ language skills during interviews, using 1–5 criteria per category. You can create multiple templates per language and choose which one is active for that language.

​
Where to find it

Sidebar → AI → Interview Evaluation (labeled Evaluation Rubrics in the product header).

​
What this page controls

  • Language proficiency scoring — not job-specific interview questions.
  • Templates grouped by language (e.g. English, Spanish, French, Portuguese, Japanese, Arabic — whatever your workspace supports in the UI).
  • For each template: proficiency categories (e.g. grammar, fluency — the exact list comes from your default categories), each with plain-language definitions for scores 1 through 5.
  • One active template per language — the active rubric is what Milo uses for that language’s proficiency evaluation until you switch it.
Per-question rubrics, weights, minimum score to pass, and follow-ups are set on each interviewer (workflow steps / questions), not on this page. See Configuring Your AI Interviewer.

​
Templates and “active”

  • New Template — Pick a language, name the template (e.g. “Customer service,” “Technical roles”), and start from the workspace default categories and score definitions.
  • Edit — Rename the template and edit the Scoring Rubric: expand each category and fill in what 1 (weakest) through 5 (strongest) mean for that category.
  • Set Active / Deactivate — Only one template can be active per language at a time. If none is active for a language, the UI shows No active template until you activate one.
  • Saving updates the template (the product may version it behind the scenes).

​
AI Tune

On each category, AI Tune lets you describe how you want criteria adjusted (e.g. stricter for senior roles, more lenient for entry-level). Milo proposes updated 1–5 text for that category; review and save when it matches what you want.

​
How this fits with overall candidate scores

  • Language proficiency rubrics here apply to how Milo evaluates language in line with your workspace standards.
  • Role-specific answers are still judged using evaluation criteria on each question when you build voice/video (and similar) steps — see How Scoring Works for the 1–5 anchor model and how mid scores are inferred.
  • Hard requirements (yes/no, dealbreakers) use Structured Questions, not this rubric screen.

​
Short tips

  • Write 1 and 5 so a recruiter could apply them consistently; the middle scores follow from that scale.
  • Use separate templates when bar levels differ (e.g. support vs. executive comms) and switch the active template per hiring program if needed.
  • After go-live, spot-check a few completed interviews; if language scores feel high or low, tighten or loosen the category text or use AI Tune.

​
FAQ

  • Is this the same as question weights on my posting? No. Weights and per-question rubrics live on the interviewer / posting setup. This page is language proficiency templates only.
  • Can I have different rubrics for the same language? Yes — multiple templates per language; only one is active at a time.
  • Where do I set “minimum score to pass”? On the interviewer configuration for the relevant step (e.g. voice/video questions), not on Interview Evaluation.

​
Next steps