Invitation: Dump Your Dashboard – or not?
How to talk to your business data with AI and know when to trust the answer.
LIVE on Wed, Sep 2 · 4–5pm CET
Imagine using AI to ask your business data:
“Which customers are growing fastest?”
“What was revenue last week?”
“Can you break this down by region?”
You’ll get an answer immediately, without waiting for an analyst to build another report or dashboard.
AI tools can already query business data this way, today. In fact, every BI vendor is selling you a “talk to your data” feature right now.
Turns out: giving an answer isn’t difficult. The harder question is whether the answer is actually right. Especially when your company has its own definitions, business logic, and edge cases.
In this live workshop, we’ll test it properly — and you’ll understand what it takes to make “Chat with your data” work in a demo vs. in production.
I’ll connect ChatGPT to a real cloud data warehouse and ask it real business questions. Then we’ll compare its answers with ground truth. Some will be correct. Others will go wrong in ways that look completely reasonable at first.
We’ll trace those errors, fix what we can, and keep pushing until we find the point where AI stops being useful.
By the end, you’ll know which business questions you can realistically answer with AI, where you need guardrails, and which questions should stay in your dashboards.
What we’ll test live
We’ll get into questions like:
Can ChatGPT really query your company data directly?
How difficult is it to connect AI to an existing data warehouse?
What happens when you ask AI a question that already exists on your dashboard?
Why can two systems using the same underlying data return different answers?
What happens when your company has its own definition of “revenue”?
How do you stop AI from confidently choosing the wrong interpretation?
Which business questions work surprisingly well?
Which seemingly simple questions are much harder than they look?
When can business users answer questions themselves – and when should the request still go to an analyst?
Can AI reduce the endless requests for new pivots, cuts, and one-off reports?
Are there situations in which a dashboard is still the better tool? (Spoiler: yes!)
What does a useful setup actually cost?
Where does this genuinely save enough work to earn its place in your business?
The goal is to help you judge where AI analytics is genuinely useful in your own environment, and where you still want the certainty of a governed dashboard or an analyst. Plus, we'll give you some pointers on how to actually set up a “talk to your data” system (and avoid the beginner mistakes).
What this looks like
Both are working from the same underlying data. During the workshop, we’ll trace the discrepancy and see exactly where the errors happened – and how to resolve them.
Wed, Sep 2
4–5pm CET
Live online + recording
Recording included.