Who matters?
Tell Cocoon which people should always be able to reach you.
How Cocoon works
It starts with a conversation.
When you first meet Cocoon, it asks you to name your personal AI representative. Then you talk. You explain the people who matter, the calls and messages you would rather not handle and the way you want everyday situations dealt with.
There are no complicated dashboards to configure. Cocoon asks questions when it needs clarity, learns naturally through conversation and becomes more useful over time.
The simple version
How does Cocoon work?
Cocoon starts by having a conversation with you.
You name your representative and explain how you want common situations handled. Cocoon asks questions, learns your preferences and confirms the boundaries it must work within.
You do not need to teach Cocoon your whole life before it can help. Start with one thing you would like it to handle, then add people, preferences and permissions naturally over time.
Name your representative. Have a conversation. Set the boundaries. Cocoon takes it from there.
Meet your representative
Your Cocoon begins with a name and a conversation.
The first thing Cocoon asks is what you would like to call your representative. That is the representative that answers, checks, follows up and handles approved interactions for you.
You then have a simple conversation about your life and how you want things handled.
What can wait?
Explain which calls, messages and routine requests you would rather not handle yourself.
When are you unavailable?
Set the moments when your attention should be protected.
What may it handle?
Agree which appointments, messages or tasks can be dealt with without asking each time.
What must come back?
Choose the decisions and situations that always need your involvement.
What is off limits?
Make clear what Cocoon should never do without permission.
You do not configure Cocoon. You get to know each other. You can begin with a few useful instructions and continue the conversation as you use it.
From interruption to outcome
Five steps. One clear purpose.
Every interaction follows the same principle: understand the situation, respect the boundaries and protect the person.
- 01Get to know youStart with a conversation
- 02Learn what mattersPeople, routines and priorities
- 03Agree the boundariesSet what it may handle
- 04Represent youHandle approved interactions
- 05Bring back the outcomeReturn what needs you
01
Get to know you
Cocoon begins with a conversation.
You name your representative and talk about the parts of everyday life where help would make the biggest difference.
You can start with one simple instruction. Cocoon asks questions when it needs clarity and learns more naturally as you use it.
Simple to begin. More useful over time.
02
Learn what matters
Your representative learns your people, routines and priorities.
It gradually understands who should always get through, when you are unavailable and how you prefer routine situations to be handled.
Better context means fewer assumptions.03
Agree the boundaries
You decide what Cocoon may handle.
Set what it may do, what it must ask and which people or situations should always come directly to you.
Permissions can be simple to begin and become more specific whenever you choose.
Useful because the boundaries are clear.04
Represent you
Cocoon handles approved interactions in your way.
It can answer, ask questions, gather information, confirm routine details or complete an action within the boundaries you have agreed.
It acts for you without pretending to be you.
05
Bring back the outcome
You receive the useful result, not the whole interruption.
Your representative returns what happened, what was agreed and anything that still needs your decision.
Less noise. More useful context.One ordinary moment
A call arrives while life is already happening.
You are helping your child with homework when an unknown number calls. Without Cocoon, you stop, ignore it and wonder, or deal with it later.
The call is received
Cocoon answers and explains that it is handling the call on your behalf.
The reason is understood
The caller is from a dental practice and needs to move an appointment.
Your permission is checked
Cocoon knows the appointment may be moved within certain days and times.
The interaction is resolved
A suitable alternative is agreed.
The result comes back to you
You receive a short summary with the new appointment and the option to change it.
The call was handled. The important moment continued.
Explore the AI Call Assistant
Personal context
Cocoon becomes useful by understanding what matters to you.
A representative cannot act responsibly without context. Cocoon is designed to build that understanding gradually, with your involvement.
Your people
Who matters, how they relate to you and who should be able to reach you.
Your availability
When you are working, resting, travelling, with family or available for interruptions.
Your preferences
How you like routine situations handled and which options you normally choose.
Your responsibilities
The appointments, commitments and areas of life you need help managing.
Your knowledge
Information Cocoon may use when answering an approved question or completing a task.
Your boundaries
What Cocoon may do, what it must ask and what it should never handle.
This context should remain visible and adjustable. You should be able to review what Cocoon understands, correct it and change your permissions over time.
Control at every stage
Cocoon acts because you allowed it to, not because it assumed it could.
Cocoon is designed around permission rather than unrestricted autonomy. You decide where its role begins and ends.
- Screen unknown calls
- Handle approved WhatsApp, SMS and connected messages
- Confirm routine appointments
- Share limited availability
- Reschedule within approved times
- Escalate urgent people or situations
Your rules may change according to the person, time of day, location, task, information requested or level of consequence.
Cocoon should never turn convenience into loss of control.Explore Privacy & Control
Good representation knows when to stop
Some moments should always come back to you.
Cocoon should not force every interaction toward automation. It should recognise when the situation is emotional, consequential, unclear, outside permission or simply better handled by you.
- The caller asks for something outside permission
- The person wants to speak with you directly
- A human relationship matters more than efficiency
- You have asked to be involved
The purpose is not to remove you from your life. It is to protect your place within it.
Across everyday life
One representative across calls and connected channels.
The same personal understanding can support calls, WhatsApp, SMS, Telegram, Facebook Messenger, Instagram messaging, appointments, availability and routine life administration.
Calls
Screening, answering, gathering information and returning clear summaries.
Appointments
Checking, confirming, rearranging and reminding.
Messages & channels
Handling approved WhatsApp, SMS and connected conversations while escalating only what matters.
Life administration
Following up, checking status and helping organise small recurring tasks.
Availability
Understanding when you can be reached and protecting the times you cannot.
Personal coordination
Helping trusted people understand what is happening without making you repeat everything.
The scope of Cocoon will grow carefully. Every capability works through the same clear pattern: understand, ask permission, act within the boundary and return the outcome.
A different kind of AI
Cocoon is not designed to become another thing you have to manage.
Not a chatbot
You should not need to open a window and explain every situation from the beginning.
Not a business dashboard
Cocoon is built around everyday life, not sales pipelines, team metrics or corporate workflows.
Not an autonomous decision-maker
It should not make consequential decisions without your knowledge or permission.
Not a replacement for relationships
It should protect human connection, not automate it away.
Not a constant interruption
The product itself should remain quiet unless it has something useful to return.
Cocoon should feel less like software you operate and more like a calm layer around your life.
Questions, answered
How Cocoon works in practice.
What does Cocoon learn about me?
Cocoon may learn the personal context you choose to provide, such as important relationships, routines, preferences, availability and permissions. You should be able to review and change that information.
Does Cocoon make decisions for me?
Cocoon may make low-risk decisions within boundaries you have already approved. Anything outside those boundaries should come back to you.
Can Cocoon handle messages as well as calls?
Yes. Where channels are connected, the same representative can work across approved calls, WhatsApp, SMS, Telegram, Facebook Messenger and Instagram messaging while using the same relationship and permission context.
Can Cocoon answer calls for me?
The AI Call Assistant can answer approved calls, understand why someone is calling, gather information and return a summary or escalate the call when needed.
Will callers know they are speaking with AI?
Cocoon makes it clear when an AI is handling an interaction. Representation should not rely on misleading someone into believing they are speaking directly with you.
Can I change what Cocoon is allowed to do?
Yes. Permission and control should be adjustable as your preferences and circumstances change.
What happens when Cocoon is unsure?
It should ask for clarification, pause the interaction or return the decision to you rather than making an unsafe assumption.
Your life remains yours
Cocoon handles what it can. You stay in control of what matters.
The best personal AI should not ask for your attention all day. It should understand your boundaries, protect your time and quietly return when there is something worth knowing.