Ask your meeting: conversational inquiry

Traditional full-vault AI search often suffers from fuzzy vector retrieval and hallucinations. Oyma adopts a scoped, deterministic approach: the Ask Workbench grounds conversational questions specifically in the transcript and Markdown notes of your active document.

The conversational inquiry workflow

1. Open any meeting note or active recording in Oyma.
2. Select the Ask tab in the right inspector panel (or press ⌥⌘A).
3. Type questions like:
   • “What budget number did David commit to for Q4?”
   • “Summarize the technical objections raised regarding GraphQL.”
   • “List every action item assigned to Sarah.”
4. The model responds with concise answers accompanied by clickable timestamp links.

Ask Workbench
Q: What was the agreed latency SLA?

A: The team agreed to a p99 latency SLA of 150ms for
the ingestion endpoint ([14:28]). Alex noted that exceeding
this would breach customer contracts ([15:02]).

Clickable timestamp citations

Every statement in an Ask response includes citation badges (e.g. [14:28]). Clicking a badge:

• Jumps the transcript view directly to that exact second.
• Highlights the corresponding spoken sentence.
• Seeks the audio player to that timestamp so you can verify the spoken words yourself.

Prompt chips for rapid triage

The Ask workbench includes one-click prompt chips for rapid meeting debriefing:

• Decisions: “List every binding decision confirmed on this call.”
• Action Items: “Extract all deliverables with owner names and deadlines.”
• Follow-Ups: “Identify open questions that were left unanswered.”

Related guides

• Summary generation in AI meeting summary archetypes.
• Local Ollama setup in Ollama local setup and model configuration.
• Cost controls in Token usage and cost controls.