Maya Chen
VP Sales · Northwind Analytics
Security
Wino AI acts on your LinkedIn and email, so it is built to stay inside the lines you draw. Here is how we keep your workspace private and your accounts healthy.
Access and isolation
Data is organised by workspace, the app runs on your signed-in session, and your channel credentials never pass through us as passwords.
Your prospects, messages and learnings are organised by workspace, and Row Level Security is enabled on database tables so data stays with the workspace it belongs to.
The app runs on your signed-in session, and backend functions check that session before they act on your workspace.
LinkedIn, Gmail and Outlook connect through a secure hosted sign-in flow. You never paste your LinkedIn password into Wino AI.
Invite teammates as Owner, Admin, Member or Viewer, so each person gets the access their job needs.
Human in the loop
Agents draft. You decide. Approval rules set what can auto-send, and everything else waits in your queue with the agent’s reasoning beside it.
Maya Chen
VP Sales · Northwind Analytics
Sending safety
Wino AI paces outreach the way a careful person would. You set the limits; the agents stay under them.
New leads get profile views and post likes first. A connection request waits for at least 2 touches over 2 days, or 5 days at most.
Limits on connection requests, messages and profile visits, with live counts of what is used and what is left.
Agents work inside your hours and time zone, and the scheduler coordinates across your connected accounts.
Follow-ups keep at least 48 hours between direct messages, and meeting asks wait for the right phase.
Today's limits
Resets in 6h8 left today
22 left today
50 left today
Data handling
The learning loop is powerful, so it comes with guardrails.
Prospect data you import or find (LinkedIn profiles and posts, Apollo contact and company data, CSV uploads), your company profile, and the messages sent and received on your connected channels.
Learnings from your outcomes are cleaned before they are added to an agent’s instructions, and they stay with your workspace.
Learning can be switched off at any time. Agents then fall back to their default rules.
Learnings lose confidence over time and are removed once they are no longer reliable.
If you think you have found a vulnerability, or you need details for a security review, email us. Please include steps to reproduce and give us a chance to fix the issue before sharing it.
Set your rules and limits once. Wino AI stays inside them.