Prepare the source
Open a recording in ScreenApp or upload a supported file. Wait for processing and check that the transcript covers the speech you need summarized. A summary cannot reliably recover words missing from the captured audio.
Attach the recording in Ask AI. State who the summary is for and what they need from it.
Ask for a specific summary
For a lecture, try: “Summarize this lecture into concepts, explanations and examples. Keep the lecturer’s qualifications. Include source timestamps where available.”
For a meeting, try: “List the decisions, actions and unresolved questions. Include owners and deadlines only when the speakers agreed to them.”
These are example instructions, not measured results. Adjust the structure to the source instead of forcing every recording into an action-item list.
Check the output
- Open the passages supporting the main conclusions.
- Check names, numbers and dates.
- Look for missing exceptions or opposing views.
- Correct transcript errors that affect the summary.
- Ask for a revised document and review it again.
If the answer is too broad, narrow the question to one topic or segment. If a conclusion lacks support, ask for the source passage or remove it.
Export or share
Generated documents support Markdown, TXT, PDF, DOCX and PPTX. Transcript export is separate and is useful when your reader needs the spoken wording rather than a summary.
Check document sharing before distributing the result. For slides, use PowerPoint creation with the reviewed source.
Understand the limitations
An AI summary selects and rewrites material. It can omit detail or misinterpret a speaker even when the transcript is mostly correct. Use Accuracy and Methodology to understand what the benchmarks do and do not establish.
If processing fails, see summary troubleshooting.