Claude Onboarding Portal
Our new Intelligence Engine — powering the next generation of Melinda Maria.
For the first time, every corner of our business — warehouse, wholesale, DTC, support, social, and our creative library — speaks to Claude in plain English. No logins. No exports. Just ask, and act.
Melinda Maria × Claude
Claude is wired straight into the tools that run Melinda Maria — predict faster, decide smarter, move first.
See it for your role
Pick your team and watch the connectors light up. These are real prompts you can paste into Claude today — from quick lookups to custom apps that run on triggers and ping Slack on their own. Scroll each list for more.
Where to actually work
Claude shows up in three places and they are not interchangeable. Most people default to Chat because it's familiar — and then do their real work the hard way. Here's the honest difference, and why Cowork, inside a project, is where nearly all Melinda Maria work belongs.
One question, one answer. You ask, Claude replies, you copy the result somewhere else yourself.
You describe an outcome, not a question — then step away and come back to finished work. Claude breaks the job into steps and runs them for you.
A terminal tool for people working inside a codebase. Enormously powerful, and not where the rest of us should be living.
And do it inside a co-work project
A co-work project is a persistent workspace with its own files, instructions, and memory. It's the difference between a tool you re-explain yourself to every time and one that already knows how we work — and it's what lets the rest of the team build on what you started.
Your instructions and its memory persist. You stop re-explaining who Melinda Maria is, which warehouse is which, and how we like things written.
Connect a folder once and it's there for every task. No re-uploading the same spreadsheet for the fifth time.
Each task starts from everything the last one learned. Week three in a project is dramatically better than day one in a fresh chat.
Decisions, drafts, and half-finished work stay in one place instead of scattered across chats you'll never find again.
Start on your desktop, check progress from your phone, come back tomorrow. The work keeps running without you babysitting it.
Anything you build in a project can go to the whole team with one link — and it opens filled with their data, through their own sign-ins. Nobody sees more than they should, and nobody waits on you for a refresh.
Because the project holds the context, a teammate picking it up gets everything you'd otherwise have to explain in a 30-minute call. Chat can't hand anything off — it lives in your window and dies there.
Tasks keep running after you close the laptop and can land in Slack or your inbox on a schedule. Your work overlaps with everyone else's instead of blocking on you being at your desk.
Go further
The prompts above ask questions. These ask Claude to build something that keeps working after you walk away — and they're the ones that change how a week feels. Every one is a real thing people here have shipped, so click any tile to see exactly what it produced, built from our own live data.
A page that re-pulls its own numbers every time you open it. Name the connector and say “live artifact” — that's the whole trick.
Work that runs on its own and lands in your inbox or Slack. Say the day, the time, and what you want — Claude sets up the schedule.
A real tool with its own web address, behind our team login, for people who'll never open Claude. This site and the roadmap board were both built this way.
One ask that spans several systems and ends in a draft you approve. Claude never sends anything without showing you first.
Claude puts a refresh control in the artifact's header, but people don't find it. Ask for one inside the dashboard and say where — then anyone you share it with can pull current numbers without asking you how.
A dashboard nobody opens twice is a wasted ask. Lead with one headline number, keep it to five or six things, colour only the exceptions, and always show when the data was pulled. Say the brand and Claude handles the rest.
Start here
Three areas, plus this overview. Jump into the connectors your team uses most, get yourself set up, or dip into the brand docs.
Everything Claude can read or do — the connectors we custom-built for our own systems, the ready-made ones, and the skills it can run.
The four-minute checklist. Switch on the systems you need, in order, signing in once each — then never think about it again.
The brand & knowledge library that teaches Claude who we are — voice, design system, product knowledge, customer — plus the source data behind it.
