Cultivated AI · Mission Control · what to bring
Mission Control builds with your real data, so the setup is heavier than a normal workshop and most of it is about access. Sort it a week ahead. If the data isn't reachable, Session 1 turns into a logins meeting.
Decisions get made live (warehouse choice, metric definitions, owners). The people who can make them have to be present.
The people who actually need answers from the data to make decisions. This is built for them.
The analyst, BI lead, or engineer who maintains Power BI and pulls reports today. They make the build real and can arrange access.
A decision-maker who can approve a warehouse direction and name owners, so the roadmap in Session 4 is real, not a wish list.
The one thing that, if missing, stalls everything. Read access is enough; we're not changing anything in your systems.
Access to the existing reports and the underlying datasets / semantic models, not just the published dashboards.
A way to get the city, store, and platform analytics out (an export, an API key, or a sample file). Bring a recent sample to Session 1.
Recent samples of whatever you want to plug in (spreadsheets, ERP exports). Anonymise anything sensitive first.
Someone who can grant access or approve a connection during the session if we hit a wall. A name and a Slack away is enough.
Mac or Windows. Chrome installed for the querying and AEO exercises.
Paid is ideal for the heavier querying, the free tier works. We'll also use a couple of assistants side by side for the AI-answer exercise in Session 4.
If Copilot in Fabric / Power BI is already on your plan, have it enabled. It's the fastest way to try the chat layer on existing data. Not required; we can demo it.
Especially the five questions you'd ask your data every week, and your one high-impact problem.
A few recent rows from Power BI, GobbleCube, and the production / supply-chain files, so we can map them for real in the room.
What ChatGPT or Gemini said when you asked "best [your category] in India." A screenshot is perfect.
You don't need a warehouse already built, a data scientist, or any new software bought in advance. You don't need to have cleaned your data first; finding the mess is part of Session 1. Bring the access and the real questions. We build the rest together.