The report you needed was a sentence all along.
No exports, no pivot tables, no waiting for someone to pull the numbers. Ask in plain English and your assistant reads your Youda directly — then charts it, summarises it, or builds the deck.
Headcount, joiners and leavers, right to work, journeys, absences, notes — and your own tasks and chats. One chat box.
Assign tasks, send a message, tick off your to-do list, update a phone number. Only what your role allows, only what you approved, and every change is logged.
Your assistant signs in as you. Team members get their own tasks and chats; managers see manager-level answers; head office sees more. The same rules as everywhere else in Youda.
The night before the board meeting.
Julia needs the people numbers, and they live in five places. She’s already in her AI assistant, so she asks there.
Youda answers, then turns the answer into four tasks for the right managers.
From leavers to lessons in one ask.
Your exit-interview agents already capture why people leave. Your assistant pulls the leavers, reads the interviews, and turns them into the deck your Monday meeting actually needed.
Themes, quotes, would-rehire — assembled in seconds, not an afternoon.
Why people really leave.
Every leaver in Youda carries a reason, a category and a would-rehire flag. Ask for the matrix and see the pattern — by quarter, by site, by role.
The difference between “turnover is up” and knowing exactly what to fix.
Small jobs, done in the chat.
A new phone number shouldn’t be a ticket. Tell your assistant, confirm the person, and it’s done — updated in Youda, synced to your HRIS.
And because it’s people data, the change lands on the profile event log — on the employee’s timeline and yours — the moment it happens.
Your whole people platform. One chat box.
Look anyone up by name, role or site — and get the profile fields you're allowed to see.
Live headcount, sliced by site or role, straight from your record.
Everyone who started or left in any window — with their sites, roles and dates.
Reasons, categories and would-rehire on every leaver — the matrix behind your retention story.
Every current check with its status and expiry date, ready to chart or chase.
Exit interviews, check-ins and agent summaries filed on a profile — readable in seconds.
Birthdays and work anniversaries coming up, so no one gets missed.
Your saved groups — first aiders, keyholders, new starters — and exactly who's in them.
Every location you operate, so questions can be sliced the way you run the business.
How cohorts are moving through onboarding, induction and reviews — who's done, who's stuck.
Change a phone number, email or emergency contact from the conversation. Synced to your HRIS, logged on the profile.
Everyone gets this one. See what's waiting on you in Youda, tick things off, add a note — without opening the app.
Catch up on the chats you're in and reply as yourself. It reads only conversations you're a member of.
Turn an answer into action: a to-do for each person, with a due date, landing in their Youda.
Send an in-app message from your organisation's assistant, or a branded email to a group — drafted in the chat, sent when you say so.
The records your agents keep about people — uniform issued, certificates, training — ready to query.
Sickness and holiday across any window, by site or by person.
Describe a survey or a workflow and get a draft agent back. Install it when you're happy — nothing goes live until you say.
People data deserves a paper trail.
Two minutes, no IT project.
In your assistant's connector settings, add a custom connector and paste in mcp.youda.co/mcp. In Claude that's Settings → Connectors.
Your assistant sends you to Youda to sign in — your own login, at your own organisation. No shared accounts, no API keys.
You see exactly what your assistant will be able to do before you approve it. Change your mind later? Remove the connector and access ends.
The whole thing in a minute and a half: connect, sign in, allow, then ask.
Questions people ask
What can my assistant see?
Exactly what you can see, and no more. Every answer is filtered to your access level — a team member gets their own tasks and chats, a site manager gets manager-level answers, head office gets head-office answers. Sensitive fields like pay never pass through unless your role allows them.
Can it change things?
Only the things you could change yourself in Youda, and only the kinds of change you allowed when you connected: your own tasks and chat replies for everyone; assigning tasks and sending messages for managers; contact details, profile notes and installing agents for head office. Changes to a person's record land on their profile event log, attributed to whoever asked.
Is there an audit trail?
Two. Every single request your assistant makes to Youda is logged with who, what and when. And any change it makes appears on the profile event log of both the person changed and the person who asked — the same place your team already looks.
Do I need to be an admin?
No. Anyone with a Youda login can connect. What you get depends on your role: everyone can work with their own to-do list and conversations and look up colleagues; managers and head office see and do more.
Which AI assistants does it work with?
It's built on the Model Context Protocol, the open standard for connecting assistants to other software. Any assistant that supports custom connectors with sign-in can add it. Claude is the one we test against every day, on claude.ai and the Claude desktop app.
See your own data answer back.
Book a demo and we'll connect an assistant to a Youda workspace live — ask it whatever you'd ask your People Team.
