Clairvoyance AI

Your agentic meeting copilot that instantly orchestrates workflows - while the meeting’s still going on.

Context
My Role: Founding/Solo Designer (8 weeks concept-to-design)

Problem

Meetings drive decisions but often kill momentum. Follow-ups get delayed, workflows fragment, and key next steps vanish in documentation. Most tools focus on recording meetings, not taking action.

Solution

Clairvoyance executes tasks using agents that listen to meetings, understand context in real-time, and launch workflows with minimal disruption—connecting decisions directly to action.

✅ 30–50% acceleration in post-meeting task initiation time
✅ 20% improvement in meeting-to-output conversion
✅ 50% reduction in number of post-meeting coordination tools

Research & Discovery

Research & Discovery
AI needs to provide users with just the right amount of information and controls as context changes rapidly.
✅ Users need visibility into what AI is doing without it being overwhelming. Transparent but not burdensome—pre-filled prompts, editable goals.
✅ Post-meeting clarity was key: users wanted to trace cause and action, not just notes.
✅ Facilitators wanted control; users need thumbs, edits, and insight into reasoning.
Initial discovery revealed the need to streamline touchpoints before, during, and after meetings to accelerate productivity.
✅ Pre-Meeting: “When I'm in a strategic sync, I want to move from decision to action without having to follow up manually afterward.”
✅ During Meeting: “When I'm leading a team meeting, I want to keep momentum without jumping into a second tool to log everything.”
✅ After Meeting: “When I'm wrapping a meeting, I want to know that the important outcomes were captured, assigned, and shared.”
Agentic Workflow
Before Meeting
✅ User Actions
  • Ground the AI in the meeting’s purpose and participants
  • Pre-load relevant documents, context, and past decisions
  • Confirm goals and get a head start on task readiness
🧠 AI Action: Context Grounding
User inputs (title, attendees, context, goals) activate relevant embeddings and load task-relevant agents. This primes the system with past decisions and workflows.
During Meeting
✅ User Actions
  • Detect key decisions or patterns in real time
  • Propose agent-driven workflows (e.g., create tasks, draft plans)
  • Preview actions without pulling attention from the conversation
  • Enable minimal-effort confirmations
🧠 AI Action: Interpret → Orchestrate
A listening agent monitors the transcript, detects moments of intent, and triggers workflows with user approval and minimal disruption.
After Meeting
✅ User Actions
  • Summarize what happened and what was triggered
  • Show task/log traceability to transcript
  • Rerun workflows or extend actions easily
  • Share structured output with the team (Slack, Notion, etc.)
🧠 AI Action: Agent Logs & Traceability
The system logs what was executed, by whom, and why—enabling users to review, rerun, or extend any workflow via agent cards and execution trails.

Agentic Workflow

Key Feature 1: (Pre-Meeting) Context-Aware Agent Setup

Key Features & Impact: Pre-Meeting Context Grounding

✨ Key Features

🧠 Suggested goals from calendar and past meetings
🔄 Auto-pulled context from docs and prior threads
🎯 Frictionless setup mirrors existing flow
📝 Smart agenda hints inferred from goal input

📈 Impact

✅ 2× increase in agent accuracy
✅ 25% boost in relevance of in-meeting suggestions
✅ Reduced effort with no onboarding required
Before Meeting: Context-Aware Setup

Problem

Users set up meetings with minimal context. Asking for more disrupts their flow—leaving the agent under-informed and less effective.

Solution

A lightweight setup screen adds context (goals, recent decisions, docs) without extra effort—using suggested prompts and passive signals to ground the agent before the meeting starts.

Key Feature 2: (During Meeting) Live Copilot Orchestration

During Meeting: Copilot Execution & Real-Time Workflow

Problem

Once the meeting starts, users want AI assistance that reacts in real time—but without interruptions. The agent must understand context, suggest relevant workflows, and keep track of actions transparently while users stay focused on the discussion.

Solution

A context-aware Copilot listens in and proactively suggests actions (e.g. summaries, reminders, tool integrations) via the chat panel. Users can accept, skip, or edit with one click. Confirmed workflows run live, with real-time progress updates and an easy-to-check Agent Log view.

Key Features & Impact: Live Copilot Execution

✨ Key Features

🧠 Live agent listening parses real-time conversation for intent
💬 Inline prompts suggest actions in chat (accept / skip / edit)
⚡ Instant execution of approved workflows during meeting
📋 Agent Log shows what was triggered, when, and why
🪄 Contextual awareness tied to pre-meeting inputs + live signals

📈 Impact

✅ 40% reduction in post-meeting coordination
✅ 50–70% fewer missed follow-ups
✅ Sub-2s latency for most confirmations
✅ Greater transparency through traceable agent log

Key Feature 3: (Post-Meeting) Workflow Review & Validation

Key Features & Impact: Workflow Transparency & Rerun

✨ Key Features

📋 Execution summary shows reasoning, outputs, and links
🔗 Connected tools like Jira and Notion surfaced directly
🔁 Rerun workflows with 1-click retry or validation
📂 Agent Log available for full traceability

📈 Impact

✅ 2× increase in follow-through on AI actions
✅ 20% boost in meeting-to-follow-up conversion
✅ 40% lift in trust from visible AI reasoning
After Meeting: Post-Meeting Workflow Review & Validation

Problem

After meetings, users often lose visibility into what the agent did. Without a clear way to trace results across tools, steps, and agents, they hesitate to trust or act on automated outcomes.

Solution

A workflow summary view shows each agent action as a status card with outputs, reasoning, and linked destinations (e.g., Jira, Notion). Users can trace steps, validate results, or rerun actions—without digging through transcripts.