Ops wants a five-person pilot before the Q3 pipeline review.
Call 18:42Private context into action
Never lose a customer follow-up again.
Mew is an AI follow-up agent for founder-led sales and customer teams. After a call, it connects permitted context from Gmail, Slack, docs, transcripts, HubSpot, and Salesforce, then drafts the follow-up email, customer note, and reminder with sources and human approval.
See the workflow
Join customer follow-up beta
After you apply, we review your workflow and reach out if you fit the first customer-follow-up beta.
Now recruiting founder-led sales and customer teams with real follow-up pain.
Follow-up email, customer note, reminder, risks, and missing info after each call.
Per founder or customer-facing seat while we validate weekly usage and approvals.
Teams who live across calls and tools like Gmail, Slack, docs, HubSpot, Salesforce, and spreadsheets.
The first product is one painful loop: customer call to approved follow-up.
Mew starts narrow: it turns scattered customer context into the packet a person usually has to reconstruct by hand.
A customer call ends. Context is already scattered.
Stephanie asked for security docs. Luis wants ROI proof. Priya needs retention answers. Sam was mentioned as rollout owner. The customer record is still stale.
Mew pulls together the permitted sources, marks uncertainty, and prepares the next action packet for review.
Search the transcript, Gmail, Slack, docs, calendar, and CRM before rewriting the follow-up, updating fields manually, and remembering to nudge later.
Review one source-backed packet: email draft, customer note, stale-field changes, risk flags, missing info, and no-reply reminder.
Follow-up packet
Email draft
Customer note
Reminder
References the pilot scope, security packet, ROI question, and Friday review ask.
Sources: Call 18:42 · Gmail Tue · Slack from MayaLogs champion, finance buyer, security reviewer, next step, risks, and missing budget owner.
Sources: transcript · customer record · security threadIf there is no reply by Friday morning, remind the owner exactly what to send.
Source: calendar hold · no-reply monitorMew starts as the company brain for customer follow-up.
The useful answer is rarely in one place. It might be in a call transcript, an old email, a Slack promise, a customer field, a PDF, a calendar event, or a tab you had open yesterday.
The wedge is customer follow-up because the pain is frequent, expensive, and measurable. The bigger vision is private context into action for every relationship and workflow where important context gets scattered.
Mew updates a living map from conversations, tasks, and corrections.
Your map should not be a static diagram. It should refresh as real work happens, then use AI to suggest what changed, what connects, and what needs action.
Conversations update the map
Calls, messages, emails, and meeting notes add people, promises, questions, decisions, and source links.
Tasks update the map
Reminders, completed work, stale follow-ups, and corrected notes change priorities and keep the map current.
AI connects the pieces
Mew can suggest relationships, clusters, drafts, and next steps, but important claims still show sources and ask for approval.
Many Mews can connect into a workspace map.
A personal Mew can keep a private map for one person. With permission, selected parts of many maps can connect into a shared workspace graph for a team, company, class, project, or household.
Each person has their own Mew that understands their conversations, tasks, files, and corrections.
The map can update every few minutes while active, hourly, after meetings, or only when the user asks.
Approved people, docs, tasks, and decisions can link across maps without exposing everything private.
A team can see how customers, projects, owners, risks, documents, and follow-ups relate to each other.
AI can cluster related work, spot missing owners, find stale commitments, and propose the next useful action.
Private memories stay private unless the user shares a source, node, cluster, action, or map into the workspace.
The map is made of people, sources, tasks, and actions.
These are the core objects Mew connects as it builds memory.
A private companion for your computer that remembers useful context and helps you act on it.
The surrounding details that make an action correct: who said what, when it happened, why it matters, and what comes next.
The little details around a customer: who said what, what they asked for, what you promised, and what happens next.
A customer database like HubSpot or Salesforce. It tracks companies, contacts, deals, notes, and next steps.
Mew does not just say a thing. It points to the email, call moment, Slack note, or doc that supports it.
A living map for one customer relationship. Not a sci-fi brain thing, just the important history in one place.
A visual map of connected people, tasks, docs, decisions, and reminders.
The thing you do after a conversation: send the email, update the note, book the next meeting, or remind yourself later.
Mew can prepare work, but you decide before anything is sent, saved, or updated.
A way to tell Mew not to remember certain apps, sites, files, meetings, or moments.
What Mew can remember.
Not everything, and not secretly. The point is permissioned context: the parts you choose to make useful.
What was discussed, what was promised, which questions stayed open, and what needs a follow-up.
Email, Slack, texts, or team chats that explain what someone meant and what they need next.
Docs, PDFs, project notes, school notes, checklists, contracts, drafts, and research.
The pages you were using when you made a decision or found something worth saving.
Names, roles, preferences, promises, open loops, and relationship history.
Reminders, deadlines, next steps, stale follow-ups, and things you said you would do.
Work tools like Gmail, Calendar, Slack, docs, customer databases, and the Windows or Mac apps you approve.
When you fix a person, source, role, date, or draft, the memory becomes cleaner.
AI is useful when it has the right context.
Writing is not the hard part. The hard part is knowing which facts matter, where they came from, and how they connect to the next task.
Conversations and tasks come in
Calls, messages, docs, browser work, reminders, and corrections provide the raw context.
Mew updates the living map
It connects people, files, promises, open questions, deadlines, and next actions.
AI suggests what changed
It can draft a reply, update a customer note, make a reminder, or show a missing connection.
You decide what gets saved or sent
Mew shows the sources behind important claims before anything becomes part of the map.
The work starts scattered
The useful details live across calls, email, docs, tabs, customer records, messages, and private notes.
The map updates over time
Mew can refresh after meetings, on a schedule, when tasks change, or when you correct the map.
Privacy stays understandable
You choose the apps, sources, and memories Mew can use. Private parts stay private unless shared.
Actions stay human-approved
Mew can prepare the work, but the user approves before sending, saving, or updating anything important.
Mew works inside and outside work.
Start with customer follow-up because it is painful and measurable. The bigger product is broader: private memory that helps with any relationship, project, or plan where context gets scattered.
Student Mew
Connects class notes, PDFs, deadlines, browser research, and project tasks so studying is less chaotic.
Sales follow-up
After a customer call, Mew prepares the reply, customer-record note, and reminder.
Professional work
Keeps clients, docs, decisions, timelines, and promised next steps clear.
Team handoffs
Helps a project survive the move from one person to another without losing the important history.
Personal admin
Plans, forms, appointments, travel details, household tasks, and important emails stay connected.
Creative projects
Turns references, drafts, decisions, feedback, and loose ideas into a map you can keep building.
Student Mew
Connects notes, lectures, deadlines, saved tabs, and drafts so the next study session starts with context.
Same product underneath: useful memory first, action second, human control always.Customer follow-up is the first business use case.
Start with founders doing their own customer follow-up because they feel the pain immediately and will pay if it works. The broader product still makes sense to anyone who loses context across messages, calls, docs, tabs, and tasks.
One person
A founder or operator gets value after the next customer call.
One relationship
The customer map becomes the source of truth for people, promises, docs, risks, and next steps.
One team
Handoffs get cleaner because the important history is not trapped in one person's head.
One private memory
Over time, Mew becomes the private memory that helps across work, school, projects, and life admin.
The feedback loop makes Mew better.
Every time you accept, edit, dismiss, or correct Mew, it learns what matters, what tone sounds like you, which sources to trust, and when to stay quiet.
Mew noticed
A customer call ended, but your customer record and Slack promise are out of sync.
Draft the follow-up, update the customer note, and remind you Friday if no one replies?
Feedback tunes the next suggestion.
Every correction makes the memory cleaner.
Mew should feel like a map you can fix, not an AI box you have to trust blindly.
permitted context
draft or remind
your corrections
People stay clear
If Mew guesses someone's role wrong, you fix it once and the customer map improves.
Claims stay grounded
Important suggestions show a source chip or a low-confidence warning before you approve them.
Timing gets better
Useful nudges teach Mew when to speak up. Dismissed nudges teach it when to wait.
The map changes as the work changes.
Mew keeps a living graph of people, emails, docs, calls, tasks, and accepted actions. You can correct it directly, so the helper gets more useful instead of guessing forever.
People
Evidence
Customer record
Open questions
Rollout
Actions
call ended
record changed
draft approved
low confidence
Edit: mark "Stephanie" as champion, not buyer
Meeting ends
Mew notices a customer moment.
Context gathers
Relevant calls, emails, docs, and tasks connect.
Draft gets reviewed
You approve, edit, or ask for a better source.
Memory updates
The map keeps the corrected action history.
Trust is a product requirement, not a footnote.
Mew starts with explicit integrations and human-approved actions, then adds deeper computer-level context only after users trust the product.
Calendar, Gmail, customer records, Slack, and transcript imports first. Selected docs and per-app capture later.
No autonomous sending and no customer-record updates until a person reviews sources and confirms the action.
Source allowlists, private mode, retention controls, per-account deletion, and an audit log are core product defaults.
Where Mew fits.
There are great tools for notes, customer databases, and chat. Mew is the connective tissue between them: it remembers the context and helps you act on it.
Great at summarizing a call. Less useful when the answer also depends on old emails, docs, Slack, and the customer record instead of just the meeting.
Great at storing structured customer info. They get stale when the real work happens elsewhere.
Great when you ask a good question. Mew tries to have the context ready before you need to ask.
A private memory map for work and life that shows sources, drafts next steps, and learns from your corrections.
Capability
Meeting tools
Customer DB
Chatbots
Mew
Meeting transcriptYesSometimesOnly if providedYes
Email thread contextPartialPartialConnector dependentYes
Slack and doc promisesNoRareConnector dependentYes
Customer record updatesPartialYesManualApproval first
Shows where facts came fromMeeting onlyDatabase onlyPrompt dependentEvery claim
Reminder and no-reply loopLimitedTask basedManualBuilt in
Editable memory mapNoFieldsNoCore asset
How Mew handles the obvious risks.
The product needs useful memory, but it also needs boundaries, sources, approval, and trust.
Start with customer follow-up because it is concrete. Keep the larger memory vision visible but secondary.
Meeting notes know the call. Mew also uses the email thread, docs, Slack promise, and customer record around it.
A CRM stores customer data. Mew helps get the right data into it after the work actually happens.
Progressive permissions, visible capture, source allowlists, no training by default, and human-approved actions are product requirements.
Mew should show sources and mark uncertain claims before you approve anything.
Mew belongs where the work actually happens. It is being built for both Windows and Mac, with permission controls for the apps and sources each person approves.
Mew in one page.
The product direction, reduced to the essentials.
What it is
A private helper for your computer that remembers customer context across apps and turns it into source-backed next steps.
First user
Founders and small teams who handle customer calls themselves and hate rebuilding context afterward.
First product
After a call, Mew prepares a follow-up email, customer-record note, and reminder with sources.
Why now
AI can write. The missing piece is the private context around what actually happened.
Why it improves
The user's corrected memory map gets cleaner over time: trusted sources, edits, accepted actions, timing, and preferences.
Expansion
Sales follow-up first, then customer handoffs, recruiting, professional services, and other relationship-heavy work.
What is live today, and what comes next.
This keeps the ambition honest: build the narrow workflow first, then expand into broader private memory.
- Source-backed follow-up packet demo
- Editable customer memory graph
- Email, customer note, and reminder review drawer
- Google and email/password accounts
- Firestore save for signed-in map edits
- Transcript and email import loop
- HubSpot/Salesforce read and approval-first writeback
- Reply and no-reply monitoring
- Source confidence and correction memory
- Design partner onboarding
- Workspace maps from many personal Mews
- Customer success handoff memory
- Recruiting and professional-services workflows
- Personal admin, school, and creative project maps
- Windows and Mac app surfaces with permission controls
What we are trying to prove with beta users.
The beta is focused on whether people actually approve Mew's actions after real customer calls.
“I spend the hour after a customer call reconstructing what I promised. If Mew only gave me the follow-up packet, I would use it every week.”
Founder-led sales user we want in beta
“The useful part is not another summary. It is knowing what changed in the account and what I need to do next.”
Customer-facing operator persona
10 paying design partners, 100+ real post-call actions, 60%+ draft acceptance, 50%+ customer-note acceptance, and median follow-up under 30 minutes.
The current demo already proves the core surface: source-backed graph, account actions, auth, persistence, editable maps, and a review flow. The next step is connecting real sources and measuring approval.
Join the Mew beta.
We are starting with founder-led sales and customer teams that handle 5+ important calls or threads each week. The broader Mew vision is still private context into action, but the first beta is customer follow-up.
Inspect your company memory.
This is the interactive product demo. Click around, drag nodes, add or delete memories, and jump between connected graph galaxies.
Demo map
Explore the sample workspace. Create an account to edit and save your own.
Your Company Memory
A source-backed memory map with people, call notes, message threads, promises, customer-record status, questions, and next actions.
Simple view starts as dots. Click Detailed View to inspect every memory node.
Review Draft
Your company follow-up packet
Bring Mew into your next customer follow-up.