Shadow Notes: AI Meeting Intelligence
Your meetings are generating institutional knowledge — and most of it disappears. Shadow Notes captures, transcribes, and structures every meeting automatically — privately, on-device, and directly into your team's Outlook calendar. No cloud uploads. No compliance risk. No behavior change required.
The Organizational Cost of Undocumented Meetings
Every decision, commitment, and escalation made in a meeting is institutional knowledge. When it isn't captured, it doesn't just become a personal inconvenience — it becomes organizational risk.
Compliance Exposure
Undocumented decisions in regulated environments (healthcare, finance, government) create audit gaps.
Leadership Continuity
When a senior team member leaves, so does everything they held in their head from three years of meetings.
Delivery Accountability
Commitments made verbally in meetings that are never tracked lead to missed handoffs, duplicate work, and blame cycles.

Shadow Notes eliminates this gap — automatically, privately, and without changing how anyone runs their meetings.
nvisia AI Lab · Prototype
Shadow Notes
Built by Mark Panthofer
A Windows desktop application that listens to your meetings, transcribes them locally, summarizes them on-device, and quietly places a structured note right next to the meeting on your calendar.
No Cloud Recording
Audio never leaves your device during processing
On-Device AI
Transcription and summarization run locally on your GPU
Auto-Organized
Notes placed directly on your Outlook calendar automatically

The recording is destroyed when it's done. Only the summary remains.
The Meeting Notes Problem Nobody Talks About
The Reality of a Packed Day
14 meetings in a day. Wildly varying context. Critical decisions made, commitments given, things that need following up.
  • Typing notes is distracting and pulls you out of the conversation
  • Recording and uploading audio raises privacy and compliance concerns
  • Even when you do take notes — what do you actually do with them?

The compounding problem:
It's not one missed note — it's the accumulation across a team of 50 people, across 12 months, across a portfolio of client engagements. The knowledge that walks out of meetings is the knowledge organizations pay senior people to generate.
"Jotting notes down continuously is distracting — it takes away from the conversation. But the information that comes out of those meetings is often very important."
Mark Panthofer
14
Meetings Per Day
Average for a busy knowledge worker
0
Notes Captured
When distraction takes over
Local, Private, Automatic
Shadow Notes works entirely within your existing Microsoft 365 environment — no new tools to deploy, no user training required, no change to how meetings are run. From the moment your meeting starts to the moment your summary appears on your calendar, it works in four seamless steps.
Step 1 — Capture
Connects to your calendar via Office 365. Uses Windows audio loopback — capturing both sides of the call, not just your microphone.
Step 2 — Transcribe
Audio transcribed locally using Whisper on your machine's GPU. The audio file never leaves your device.
Step 3 — Summarize
A 3.59B parameter model running locally captures key decisions, action items, and discussion points. No cloud calls. No API.
Step 4 — Store
Summary written to a "Shadow" calendar event in Outlook. Full transcript goes to OneDrive. The recording is destroyed.
Security and Privacy by Design
Most organizations are cautious about AI meeting tools because the recordings go somewhere — a vendor's cloud, a third-party service, an LLM API call with sensitive content in the payload. Shadow Notes is designed from the ground up to avoid all of that.

Designed for Regulated Industries
Most enterprise AI tools create compliance risk by routing sensitive audio and transcripts through third-party cloud infrastructure. Shadow Notes is architected to eliminate that risk entirely — making it viable in environments where conventional meeting AI tools are prohibited.
  • Financial services teams discussing non-public information
  • Healthcare organizations subject to HIPAA considerations
  • Government contractors with data residency requirements
  • Legal teams with privilege concerns
Audio Never Leaves the Device
The complete audio stream is processed entirely on your local machine — no uploads, no streaming to external services.
Transcription Runs Locally
OpenAI Whisper executes on your GPU. Your words stay on your hardware.
Summarization Runs Locally
The on-device LLM processes your transcript without any external API calls or cloud dependencies.
Recording Destroyed After Processing
Once the transcript is generated, the audio file is permanently deleted from your device.
Only Text Summary Reaches OneDrive
Within your own Microsoft 365 tenant — all data stays within drive-level encryption.
"Ethically there's no issue. It's just like taking really good notes — except your notes are automatically organized where they're supposed to be."
Mark Panthofer

All data stays within your Microsoft 365 tenant and drive-level encryption. Zero third-party exposure.
From Notes to Intelligence
The summary in your calendar is just the beginning. Because Shadow Notes has access to your Office 365 environment, it can feed into a Copilot agent that transforms raw meeting data into actionable intelligence.
"If I have the API, there's very few things I can't do."
Mark Panthofer


Weekly Commitment Review
Every commitment, decision, and action item surfaced across all meetings in a given week — reviewed in minutes, not reconstructed from memory.
Client and Project Intelligence
Meetings auto-tagged by client or project. Leaders can query the full history of what was discussed, decided, and promised for any engagement.
Time Tracking and Billing Accuracy
Calendar category mapping ties meeting time to client projects — creating a passive, accurate record of where senior time is actually spent.
What This Looks Like at Scale
Shadow Notes begins as a personal productivity tool. But when deployed across a leadership team or client-facing organization, it becomes something more: a searchable, structured record of institutional decision-making.
Leadership Transitions
A new VP steps into a role and queries every meeting from the previous 6 months — understanding the full context of every client relationship and internal initiative without a single knowledge transfer session.
Client Accountability
A project manager pulls every commitment made to a client over the past quarter — before a QBR — in under 60 seconds.
Regulatory Audit Readiness
A compliance officer audits meeting records for a regulatory review — structured, timestamped, and organized by topic.

This is not a note-taking app. It is an organizational memory system.
Tech Stack
For the technically curious: Shadow Notes is built entirely on local-first, Microsoft-native infrastructure — no new vendors, no new contracts, no new security review.
Shadow Notes is built on a carefully chosen set of local-first technologies that prioritize privacy, performance, and deep Microsoft 365 integration.
Platform
Windows desktop application — native performance, full GPU access
Audio Capture
Windows Loopback API — captures the complete merged audio stream of both sides of the call
Transcription
OpenAI Whisper running locally on-device GPU — no cloud, no upload
Summarization
3.59B parameter local LLM on-device GPU — zero external API calls
Calendar Integration
Microsoft Graph API / Office 365 OAuth — deep calendar and OneDrive access
Storage
OneDrive for text summary only; local drive for encrypted, transient audio — destroyed after use

Built with AI-assisted development — a fitting approach for an AI-powered tool.
Experience Shadow Notes at the nvisia AI Symposium
Visit the Shadow Notes booth to experience the full capture-to-summary workflow in action.
1
Live Capture-to-Summary Demo
Watch the complete workflow from meeting start to structured calendar note — in real time.
2
Shadow Calendar Event
See the "Shadow: [Meeting Name]" event appear right next to the real meeting on an Outlook calendar.
3
Microsoft 365 Data Boundary
Understand exactly how the system stays entirely within your Microsoft 365 tenant — no data escapes.
4
Compliance Discussion
Explore where this privacy-first approach could apply in compliance-sensitive environments.
5
The Roadmap Conversation
Explore what Shadow Notes looks like as a deployed enterprise capability — and where the nvisia AI Lab is taking it next.
Meet the Builder
Mark Panthofer will be on-site to demo and discuss Shadow Notes in depth.
Ask questions about the architecture, the privacy model, and what comes next for this prototype.
Mark brings deep experience in enterprise system architecture and AI-assisted development. Shadow Notes is one of several prototypes emerging from nvisia's AI Lab — a practice dedicated to building AI tools that solve real enterprise problems with responsible, privacy-first design.

Part of the nvisia AI Lab — where real prototypes solve real problems.
Interested in Shadow Notes for Your Organization?
Whether you're exploring AI for your enterprise, evaluating privacy-first meeting intelligence, or want to learn more about what the nvisia AI Lab is building — we'd love to connect.