Screen-aware roleplay needs a visibility receipt

Separate three session states

Represent screen sharing, AI visibility, and screen-based scoring as separate fields. For each transition, record the participant, tenant and plan, roleplay and rubric version, device and application, selected window or display, recording state, AI-view state, active Screensharing Goals, person who changed the control, event time, and confirmation shown. The participant may stop AI viewing while a screen recording or administrator-defined rubric remains active; the public documents do not resolve that relationship.

Make the current combination visible before the roleplay begins and whenever one field changes. A screen-share icon should not serve as the only notice for model observation or scoring. Use plain states such as shared and recorded, shared and AI-visible, and shared and scored, with unavailable or unverified where the platform cannot report one. Keep the receipt specific to the tested plan, tenant, application, roleplay mode, and version rather than generalizing a help-page description to every deployment.

Link feedback to the governing state

For each AI response, feedback statement, or Screensharing Goal score that refers to visible material, attach the relevant session interval, AI-view state, recording state, rubric identifier and version, source marker, model version, and review status. Label evidence derived from speech, typed interaction, scenario instructions, uploaded material, a live shared screen, or the screen recording separately. A numeric score does not reveal whether the AI evaluated the intended screen interval or whether a visibility-off period entered later processing.

Test a rubric whose required screen sequence is deliberately different from the spoken sequence. Turn AI visibility off for one labeled screen while recording continues, then inspect live responses, the saved recording, analysis, and score. Repeat with no Screensharing Goal and with screen sharing stopped entirely. The acceptance question is whether each output can be reconciled to the state that authorized its source, including an explicit unknown when the product offers no event-level lineage.

Accept the control with a transition matrix

Run a synthetic session across the meaningful combinations of sharing, AI visibility, recording, and screen-scoring configuration. Alternate the AI-view choice while different labeled slides are visible; inspect the participant notice, AI replies, saved recording, analysis, rubric score, and administrator view. Include a late toggle, rapid toggle, network interruption, page reload, an unsupported surface, accidental share, and a participant who declines AI viewing. Record observed behavior instead of assuming that the more restrictive state governs every other function.

Assign owners for the participant explanation, roleplay and rubric, platform configuration, score review, accessibility, incident handling, and change control. Keep screen-derived evidence out of certification, work access, performance, or employment decisions unless the permitted state, scoring method, human review, and appeal are separately approved. Reopen the matrix when the platform, model, device flow, roleplay, rubric, analytics, or plan changes. A bounded pilot can proceed only when reviewers can reconstruct which state governed each screen-derived output.

Turn this source into a reviewable decision

For AI Coaching Platforms for Leadership Development, use this briefing as a dated decision record rather than a substitute for the source. Preserve Yoodli Release Notes, the exact URL, the September 28, 2026 review date, the supported facts above, the editorial interpretation, the limitations, and any buyer-specific evidence. Link that record to the decisions most directly affected: Data flow and confidentiality; Safety, boundaries, and escalation; Workflow and identity integration; Accessibility, language, culture, and user control. State whether the source changes the scope, evidence requirement, control, sequence, or only the language used to describe the decision.

Before action, name the accountable owner, affected population and workflow, exact offering or configuration, source data and rights, human decision point, exception and appeal path, complete cost, expected benefit, failure and stop conditions, retained evidence, and next review date. Keep official facts, provider statements, buyer observations, representative tests, measured outcomes, editorial inferences, and unknowns visibly separate. Reopen the record when the source, offer, model, integration, data, policy, population, responsible person, or measured result changes.

Limitations and unknowns

Yoodli is the provider and source for the September 23, 2026 release-note entry, the August 27, 2026 practice page, and the August 11, 2025 Screensharing Goals page, all checked September 28, 2026. All three predate the September 27 successful-run cutoff and therefore do not establish post-cutoff news. The pages document a choice about AI viewing, screen recording during practice, and enterprise-plan screen-scoring behavior at different times, but they do not define how the newer visibility choice affects recording, later analysis, a Screensharing Goal, historical sessions, or every tenant. They do not establish current entitlement, configured defaults, supported surfaces, complete notices, event logs, failure behavior, score accuracy, accessibility, confidentiality, participant benefit, behavior change, or business outcome. Verify current documentation and contract, an authorized tenant with synthetic content, the participant and administrator interfaces, the transition matrix and output lineage, and qualified coaching, learning, HR, employee-relations, labor, accessibility, privacy, security, procurement, records, regulatory, and legal review before live or consequential use.