Practice Hub App Overview

Practice Hub

1. What it is

Practice Hub lets employees rehearse real conversations and presentations against an AI that plays the other person — a frustrated customer, a job candidate, an audience — then scores the attempt against a rubric you define. Every practice ends in a scorecard: the learner’s own words quoted and judged, with delivery signals like pace and filler words when voice is used.

Scenarios are peer-authored by default. Employees draft them; L&D curates and promotes the best; and only a promoted Official scenario can mint a real Skills & Certifications credential.

• Enablement: Practice Hub requires a licence and is off by default. An admin enables it per tenant from the Apps Marketplace.
• What it is not: Practice Hub is not a course builder — it is the rehearsal step between learning something in Training and doing it for real. Course creation and completion tracking live in Training.

2. Standing it up

1. Go to Admin → Apps Marketplace → Practice Hub and enable the app.
2. Open Practice Hub → Settings and review the defaults. The app works immediately with its defaults — the items below are the ones worth deciding up front.
3. Decide whether employees may publish scenarios to the shared library. Allow employees to publish scenarios to the Practice Library is on by default. When on, peer submissions go through an AI content check and (by default) L&D approval before they reach the library.
4. If you want managers to review and adjust AI-generated scores, confirm Let managers review and override scorecards is on (it is by default).
5. Grant access. Practice Hub respects platform Access Management — target it to departments, locations, or specific people via groups. Any user with access can browse the library and practice; managers see the Assignments and Reviews tabs for their direct reports; admins see Curation, Analytics, and Settings.
6. Create your first scenario. Go to Practice Library → New scenario (or use Describe it, AI drafts it to generate one from a prompt). Add at least one persona and two scoring criteria, then publish.
7. Confirm it worked: employees should see Practice Hub in their app list, the scenario in the Practice Library, and a Start practice button on it.

A scenario with no persona cannot be practiced — the Start practice button is disabled until at least one persona is added.

To get started quickly, the Settings page offers Load sample scenarios (pre-built examples) and the Working with AI collection — five official scenarios covering delegation, evidence checking, exception handling, escalation, and responsible approval. Both are optional.

3. How it fits together

Scenarios and the trust ladder. A scenario is the authored unit: a situation (“an angry customer calls about a late delivery”), one or more AI personas to roleplay against, and a rubric of scoring criteria. Every scenario has a visibility tier that determines how much trust the platform places in it:

• Community — the starting tier for peer-authored scenarios. Visible in the library once published.
• Curated — L&D has reviewed and promoted it. Admin-authored scenarios start here automatically.
• Official — the highest tier. Only an Official scenario can be marked Certifiable, which means a passing attempt mints a real credential in Skills & Certifications.

Example: an employee drafts a “Handling a product recall call” scenario → it lands as a Community draft → they publish it → the AI content check runs → L&D approves it into the library as Community → after it proves popular (good ratings, high pass rate), L&D promotes it to Curated → eventually to Official + Certifiable, linked to the “Crisis Communication” skill.

Attempts and scoring. When a learner starts a scenario, the app creates an attempt — the record of one practice session. The learner converses with the AI persona (by typing or by voice), then ends the session. The AI scores the attempt against every rubric criterion and produces a scorecard with an overall score. The default passing threshold is 70 (configurable per scenario, 0–100). A scored attempt is private to the learner; managers see it only through the Reviews queue, and only for their own reporting line.

Interaction modes: conversation vs delivery. Each scenario is set to one of two modes. In Conversation mode (the default), the learner has a back-and-forth with a counterpart — a customer, a colleague, a candidate. In Delivery mode, the learner holds the floor (a presentation, a pitch, a speech) and the AI persona acts as an audience — reacting, asking questions, but never taking a counterpart role.

Separately, a scenario’s modality controls whether the learner practices by typing, by voice, or by their own choice. Modality is a content decision made by the author; the tenant-level Live voice roleplay setting must also be on for voice to be available.

Assignments and campaigns. A manager can assign a specific scenario to one or more people with an optional due date. An assignment moves through four states: Assigned → Started → Passed → Canceled. A campaign wraps an assignment around an audience (a department or group) with optional recurrence — “everyone on the support team passes this scenario every 90 days.” Recurring campaigns regenerate assignments automatically at the interval.

Score appeals. A learner who disagrees with their score can raise a score appeal from their scorecard. The appeal goes to their manager’s Score appeals queue (visible alongside the Reviews tab). The manager can uphold the score, adjust it, or correct the transcript. Score appeals require manager review to be turned on — if the setting is off, there is no reviewer to route to.

Drills. After a scored practice, if the learner underperformed on a specific rubric criterion, the scorecard offers a focused drill — a short, private practice targeting just that weakness. Drills are auto-generated clones of the original scenario, tagged as private, and excluded from the Practice Library and all admin surfaces.

Recording analysis (opt-in). When Let employees upload call/meeting recordings for private AI analysis is turned on, employees can upload a real call or meeting recording. The app transcribes it and scores it against a chosen scenario’s rubric — the same scorecard as live practice. Recording analyses are private to the uploader and excluded from the manager review queue.

Skills & Certifications integration. When an Official + Certifiable scenario is linked to a certification skill, a passing attempt automatically mints an employee skill record and a certification. If the scenario has a recertification interval, the credential carries that expiry. A manager who overrides a score downward can withdraw the credential automatically.

The AI coach. Each learner has a coach profile that the AI builds over time from scored practice. The coach tracks strengths, focus areas, and a narrative of the learner’s development. It uses this profile to personalise scenario recommendations on the dashboard and in the Practice Library, and to offer a weekly practice plan (when the digest setting is on). Learners can see what the coach knows about them, set a personal practice goal, and remove any item from the coach’s memory. If Personalize AI coaching with workplace signals is on, the coach also draws on performance review themes, recognitions, and skills data.

Mobile. Employees on phones are automatically redirected to the mobile Practice Hub. The mobile experience covers the library, scenario details, and the full roleplay and scorecard flow. Manager and admin surfaces (Assignments, Reviews, Curation, Analytics, Settings) are desktop-only.

4. Running it

Curating peer submissions

1. Go to Curation. The top section shows scenarios submitted by employees, oldest first.
2. For each submission, review the scenario content and the AI content check result (advisory — your decision overrides it).
3. Choose Approve (publishes it to the library) or Return to author with a note (sends it back as a private draft with your feedback delivered to their Inbox).

Promoting scenarios

1. In Curation, the lower section lists published Community and Curated scenarios sorted by rating and usage.
2. Select one or more scenarios and choose a target tier: Curated or Official.
3. For Official scenarios, optionally check Certifiable and link at least one certification skill. A certifiable scenario with no linked certification skill cannot be saved.

Reviewing scorecards

1. Go to Reviews. The Pending tab shows scored attempts by your direct reports (or, for admins, the whole company) that have not been reviewed.
2. Open an attempt to read the full transcript and scorecard.
3. Choose Accept (endorses the AI score) or adjust the score to a number 0–100. The learner is notified of adjustments.

A review that changes a passing result to failing automatically withdraws any Training completion or certification that attempt had earned, scoped to what that specific attempt produced.

Assigning practice

1. On any scenario, click Assign. Pick one or more people (or a department), set an optional due date, and add an optional note.
2. Assignees receive an Inbox notification. The assignment appears on their dashboard.
3. Track progress on the Assignments board. Managers see assignments for their direct reports; admins see all.

By default, managers can only assign practice to their own reporting sub-tree. The setting Let any manager assign practice across the whole company widens this to all employees.

Running a campaign

1. Go to Campaigns → New campaign. Pick a scenario, an audience, a due date, and optionally a recurrence interval.
2. The campaign fans out individual assignments to every member of the audience.
3. For recurring campaigns, the next round of assignments is generated automatically at the recurrence interval.

Viewing analytics

Go to Analytics (admin only). The dashboard shows headline counters (total scenarios, published, official, total attempts, scored, passed, active practitioners), the most-practiced scenarios, and category and difficulty breakdowns. Additional panels cover assignment adoption rates, competency gaps across the team, practice by department, shelf health (active vs stale scenarios), and AI usage. An Export CSV button downloads up to 10,000 attempt records.

Handling score appeals

When a learner raises a score appeal, it appears in the Score appeals tab (visible to managers alongside Reviews). Open each appeal, read the learner’s statement and the disputed evidence, then resolve it — uphold the original score, adjust it, or correct the transcript.

5. Settings

All settings are at Practice Hub → Settings. Defaults are read from the app’s configuration schema and apply until an admin changes them.

Setting	Default	What it changes
Practice Hub AI Agent	On	Whether the in-app AI assistant can answer questions about Practice Hub.
Allow employees to publish scenarios to the Practice Library	On	Whether employees can submit scenarios to the shared library. Drafting for private practice is always allowed.
Peer scenarios need L&D approval before going live	On	When on, a peer submission goes to the Curation queue for admin approval. When off, it goes live automatically after the AI content check passes.
Let managers review and override scorecards	On	Whether the Reviews queue and Score appeals tabs appear for managers.
Let any manager assign practice across the whole company	Off	When off, managers assign only to their direct reports and their sub-tree. When on, any manager can assign to anyone.
Award Recognition points when a peer scenario is promoted	Off	Whether a promotion to Curated or Official grants Recognition points to the scenario author.
Suggest new scenarios weekly from real escalated/reopened requests	Off	When on, the AI drafts scenario suggestions weekly based on real support escalations. L&D approves them before they enter the library.
Let employees upload call/meeting recordings for private AI analysis	Off	Enables the Recording analyses tab and upload flow. Carries consent and privacy weight — two-party-consent laws may apply.
Days to keep uploaded recording files before auto-delete	30	How long raw recording files are retained. Transcripts and scorecards are kept permanently.
Personalize AI coaching with workplace signals	On	Whether the AI coach uses performance review themes, recognitions, and skills data to personalize feedback and scenario recommendations.
Send active practitioners a weekly practice plan	Off	Whether active practitioners receive a Monday Inbox digest with a personalised practice plan.
Live voice roleplay	Off	Enables speech-to-speech practice using voice minutes from the tenant’s balance. Each live session is capped at 10 minutes.

Below the toggles, the Company context section lets you describe your products, common objections, policies, and tone guidelines. The AI personas and scoring coach ground in this context when it is filled in.

6. More help

• Practice Hub FAQ — specific setup and operating questions, limits, and troubleshooting.
• Ask AI — the assistant answers questions about Practice Hub from these articles.