Today Alex shows how to use scheduled prompts in Microsoft 365 Copilot to generate a consistent daily brief: decisions, delegations, what can wait, and calendar conflicts.

Tip: Add constraints like citations and decision filters to keep the output tight and trustworthy.

• Produce a decision brief: what's agreed, what's disputed, what's missing
• Extract a checklist with owners and suggested due dates
• Draft a concise reply that asks at most two questions

Plus: a fast assumption-check prompt to stop the model from guessing.

Reusable prompt pattern:

You are my operations assistant. Read the text below. Output: a five-bullet summary; decisions that must be made with who should decide; open questions; and the next three actions to take today. Keep it under one hundred and fifty words. If a detail is not explicitly in the text, write ‘not stated.’ Text: paste the thread or document.

Prompt pattern: Turn the text below into an execution brief: decisions, commitments with owner and due date if stated, open questions, risks, and the single next message I should send. If anything is missing, ask up to three clarifying questions first.

Primary reference: Microsoft Support: Summarize an email thread with Copilot in Outlook.

• Pick the sections you actually use (decisions, next steps, short context)
• Use a reusable prompt pattern to standardize output
• Add guardrails so the AI marks uncertainty instead of inventing certainty

Host: Alex Vance

Prompt pattern included in the episode.

In this episode: how to take a meeting recap, run a second AI pass to produce a commitments register, and draft a follow-up email that asks for corrections.

Today’s episode: how to use Gmail’s email thread summaries as a starting point, then convert them into a verified decisions-and-actions brief plus a draft reply.

Host: Alex Vance

Prompt pattern: Ask for exactly four sections: top outcomes, commitments table (with confidence score), risks + clarifying questions, and three short draft replies.

Prompt patterns included: ask for an answer and a verification checklist, then convert the result into a one-page action note you can paste into an email.

Prompt pattern included: a reusable checklist prompt plus guardrails to avoid incorrect category merging.

Source: Microsoft Excel blog, Five ways to create and analyze an Excel spreadsheet with AI.

• Set five buckets that match your work
• Add guardrails for pricing, contracts, refunds, and escalations
• Reuse one copy-paste triage prompt pattern
• Force quoting to prevent made-up details

Host: Alex Vance

• The one-line context habit: The single annotation that makes AI retrieval dramatically more useful — and takes five seconds per note.
• The retrieval prompt: A structured prompt that summarizes your saved notes, surfaces gaps in your thinking, and generates questions you have not thought to ask.
• The weekly synthesis pass: A ten-minute habit that turns a passive note archive into an active thinking tool.
• Key watchout: Why setup time is the enemy of a useful knowledge base, and the rule that keeps it lightweight.

• The preparation prompt: A structured prompt that maps the other person's likely perspective, surfaces the objections you're most likely to face, and helps you distinguish the outcome you need from the one you want.
• The stress-test prompt: A single question that forces AI to argue against your position — surfacing the blind spots you've been avoiding.
• Key watchout: Why AI will be more diplomatic than the real conversation, and how to ask for a harder simulation when the stakes are high.

• The briefing prompt: A structured prompt that produces a formatted one-pager from raw pasted notes — ready to share or act on.
• The judgment layer: A second prompt that connects the briefing to a specific decision and stress-tests your options.
• Key watchout: How AI generates confident-sounding findings from thin evidence, and the thirty-second spot-check that catches it before you share.

• The constraint-first approach: Why giving AI your available hours and fixed commitments before your task list is the key to a plan that actually holds.
• The prioritization prompt: A structured prompt that identifies your top three outcomes, categorizes remaining tasks, and suggests daily focus areas by task type.
• Two watchouts: How to prevent AI from generating a sixty-hour plan, and why the plan is a starting point — not a contract.

• The structured output prompt: A ready-to-use prompt that produces an executive summary, decision log, action item table with success criteria, and a draft follow-up email.
• The refinement pass: A second prompt that converts ambiguity into clarifying questions and makes action items executable.
• Two critical watchouts: How AI invents decisions that were never made, and how to fix missing owners before the follow-up goes out.

• Template extraction workflow: How to use AI to analyze your best past replies and pull out reusable frameworks you can adapt in seconds.
• Edge case prompt pattern: A structured prompt for drafting responses to unusual or emotionally charged customer messages.
• Watchout: Where AI tone goes wrong and the one-line fix that keeps responses sounding human.

Today's episode walks through a five-step workflow that turns rough meeting notes or transcripts into a clean follow-up email and task list in under ten minutes — using any chat-based AI assistant.

• The five-step workflow: Capture raw notes, extract decisions and commitments, do a 60-second reality check, generate two versions of the follow-up (external email + internal task list), and move tasks into your actual work system.
• Reusable prompt pattern included: A ready-to-adapt prompt that outputs a recap, decision bullets, an action item table with confidence scores, and a draft follow-up email.
• Three watchouts: Indirect commitments, date hallucination, and tone drift — and how to fix each.

Sources:
Simular AI — AI Meeting Note Takers Comparison — https://www.simular.ai/alternatives/ai-meeting-note-takers

Today's episode walks through the BUILD framework: a five-part prompt structure (Background, Use case, Instructions, Length, Deliverable) that consistently produces output you can actually use — often with one or two small tweaks instead of a full rewrite.

We also cover the key mental model for knowing which tasks AI should own completely versus which ones need your judgment.

• The BUILD Framework: A five-part structure that turns vague requests into precise, usable AI outputs.
• Speed tasks vs. judgment tasks: A simple decision filter for knowing when to let AI do the full job and when to keep your hands on the wheel.

Sources:
Parker Prompts / Geeky Gadgets — Why Separating Speed Tasks Is the Secret to Mastering AI at Work — https://www.geeky-gadgets.com/ai-productivity-guide-2026/