Intelligent Meeting Recap Available Without Saving Transcript
Jul 1 2026
A significant change to the management of meeting artifacts with retention and discovery impacts
Intelligent recaps in Microsoft Teams can now be generated without retaining a transcript after the meeting. Meeting organizers can choose this setting from the meeting options before the meeting, or organizers and eligible participants can enable it during the meeting. After the meeting, the AI summary will remain available even though no recording or transcript was ever saved.
What’s Changing and When
This is a modification of an existing feature.
Microsoft 365 Roadmap ID: 558286
Preview: N/A
Expected GA: August 2026
Why This Matters
Microsoft is introducing a new “recap without saving transcript” capability in Teams that allows Copilot to generate AI meeting summaries using only live meeting context, without creating or storing a transcript or recording. The feature is available to users with Microsoft 365 Copilot (premium) licenses, is governed by existing tenant-level AI controls, and is enabled by default where Copilot is permitted. Meeting organizers (and in some cases participants) can toggle the feature before or during meetings, and no transcript or recording is retained as part of this experience.
Because these AI-generated summaries are created using ephemeral transcriptions, the underlying meeting content never becomes a governed record, and there is no primary artifact to retain, classify, or place on legal hold. As a result, organizations may find themselves relying exclusively on AI recaps without corresponding source material, raising questions about whether these recaps constitute business records, how long they should be retained, and how they should be classified. While Microsoft frames the feature as an enhancement to privacy and compliance controls, records programs may elect to clarify that those recaps are derivative content rather than authoritative records.
Key Risks Organizations Should Consider
Potential Policy Gap
This feature is explicitly designed for organizations that prefer not to retain meeting transcripts or recordings. However, avoiding retention of these artifacts does not eliminate regulatory or litigation obligations, and meeting content may still be discoverable through related emails, chats, or Copilot outputs. As a result, retention schedules and defensible disposition policies may need to address AI-generated summaries explicitly, even when the underlying source data is ephemeral.
Increased Importance of Downstream Retention Controls
Because the AI recap can be copied into OneNote, saved in a Teams chat, or forwarded via Outlook, retention enforcement may shift away from the meeting itself and toward the locations where users choose to store or reuse the recap. As a result, auto‑labeling, adaptive scopes, and location‑based retention become increasingly critical to ensure consistent governance of these derivative artifacts.
eDiscovery and Legal Hold Considerations
This feature bypasses creation of traditional meeting artifacts. A potential discovery blind spot arises if legal teams assume that the absence of a transcript or recording means no meeting content exists, even though AI-generated recaps of uncertain veracity may still influence business decisions, reflect factual statements, and be circulated or relied upon informally. From an eDiscovery perspective, organizations should not equate “no transcript” with “no discoverable information,” particularly where recaps are saved or reused in other systems. Discovery will depend on the downstream locations where Copilot content is stored rather than on Teams meeting infrastructure. As a result, legal hold scoping should focus on where Copilot-generated content is saved, typical user behavior patterns, and “non-record” meeting workflows that still produce discoverable information.
How eMerge Can Help
eMerge can help clients close policy gaps that arise from platform changes by mapping these new capabilities against existing retention, regulatory, and litigation obligations, and then updating retention schedules and defensible disposition procedures.
eMerge can also design and implement downstream retention and eDiscovery controls that reflect how employees actually use Copilot and other new Microsoft tools. That may include configuring auto‑labeling, adaptive scopes, and location‑based retention; revising legal hold playbooks and queries to target Copilot outputs; and training legal, IT, and business stakeholders on how to navigate discovery of AI artifacts. Together, these measures help ensure that derivative AI summaries are governed, preserved, and retrievable in a way that aligns with the organization’s regulatory, litigation, and records management requirements.
Bottom Line
The ability to generate intelligent meeting recaps without a companion recording or transcript may change how organizations approach information governance and eDiscovery. Consider updating administrative and end-user guidance, policies, and trainings to clarify responsibility for data and records management, validation of accuracy, and acceptable use of AI meeting recaps.
Want to talk through whether and how this new capability should be enabled in your environment?
Please contact eMerge Information Governance.