How the PokaMind layer works
What goes in, what it measures, what each agent gives back, and what is kept.
One layer under every agent
What goes in
- Meetings you record, goes to: Meeting notes.
- Practice sessions, goes to: Practice breakdown.
- Coach voice notes, goes to: A note per person.
- EMORI conversationsIn pilot, goes to: EMORI’s read.
- Your documents, goes to: Your AI assistant, through MCP.
PokaMind layer: adds context
- Tone of voice
- AccentYour own speech
- Pace
- Body languageYour own camera
- Words
Measures observable signals of how things are said
Shared as context, not measured
What comes out
Why it helps
- Know how it landed, not only what was said
- Practice before the real conversation
- Your AI assistant gets the context you share
Works with your tools
Your meetings go in, with the playbooks, policies and documents you choose to share as context. Notes and context go out to your AI assistant and your team tools.
Meetings
- Zoom
- Microsoft Teams
- Google Meet
Documents
- SharePoint
- Google Drive
AI assistants, through MCP
- Claude
- ChatGPT
- Microsoft Copilot
Where work lands
- Slack
- Salesforce
- HubSpot
Who starts it, what is kept
Records only when a person presses record
Never reads emotions or mood
Never used to hire or rank
Encrypted, in line with GDPR
Product by product
- PokaMind Notes, Web and Android
- You press record after everyone in the meeting has agreed. Each person who opts in gets feedback on their own delivery, and no one else sees it. Your notes are shared only if you choose. Speakers are labeled Speaker 1, Speaker 2 and so on. PokaMind does not recognize anyone by voice.
- PokaMind Practice, Web and Android
- A person opens a session, and the written breakdown goes to them. What a practice session records is listed below.
- PokaMind Coach, On a phone
- The coach talks for a minute after a session and checks every note before it goes out. Only the text is kept, not the voice recording.
- EMORI, In pilot, In the break room
- EMORI is our desktop robot. Staff pick a conversation to practice by saying it, and the read of their own delivery plays in their headphones or shows on EMORI's screen. Its camera is used only for the speaker's own body language, and the video is not kept. No login, so staff can stay anonymous.
- Feedback, By voice or in writing
- People choose to answer, in their own words. Tell them who will read the summary.
- Weekly digest, Every Monday
- It is written each week from the documents you share. Someone on your side approves each post and chooses who sees it.
What it measures, and what it leaves out
It measures
- Tone of voice: your pitch, volume and emphasis
- Pace and pauses
- Accent in your own English, if you turn it on
- Your body language on your own camera
- Who talks when
- Words, filler words and hedges
It is designed not to infer
- Emotions or mood
- Stress or nervousness
- Personality
- Honesty, intent or attitude
- Who someone is: no face or voice recognition
What managers see in practice
- Sees
- Who practiced and how often.
- Never sees
- Feedback, the conversation, the transcript or the written breakdown.
Your data
Session data is encrypted in storage and in transit and handled in line with GDPR. We share our security documentation and data processing terms during evaluation.
In the AI assistant you already use
PokaMind connects to assistants such as Claude, ChatGPT or Microsoft Copilot through MCP, the open standard AI assistants use to connect to tools. It gives your assistant the context of the spaces you share, and you can disconnect it at any time.
What is MCP?PokaMind