Home / Lakehouse-Analyst
Accelerator · Lakehouse-AnalystAsk your data questions in Teams or WhatsApp.
It answers in seconds with your company's numbers, respecting each person's permissions: it checks who's asking before running a query, replies with a table and its findings, and says which table each figure came from.
The agent understands your business: every table and column carries its meaning, its owner, and who can see it. That's why it answers with your definitions, not with guesses.
The situation today.
Getting answers from data is slow: it takes a specialist, access rights, and knowing SQL or the reporting tool.
Dashboards cost a license per person and answer what someone anticipated, not today's question with today's numbers.
A generic AI doesn't know your data or your business unless someone gives it that context. That's what we do: we give it your tables, your definitions, and your permissions.
What we do well.
In your channels
- Microsoft Teams and WhatsApp
- Slack, coming later
- Nothing new to install
On your platform
- Athena, Snowflake, BigQuery, or Databricks
- On top of the gold tables and their semantic layer
- Without moving your data
Governance before the query
- Table- and column-level permissions
- Your company's sign-in (Entra or Google)
- What's off-limits is explained, not shown
Answers with their source
- The figure in a table, plus two lines of findings
- The source table and its cutoff time
- The query, for anyone who wants to check it
Tested in every release
- Accuracy on real questions
- Permissions with different users
- Measured response times
Inside your pod
- Part of the monthly service
- We run it and improve it every month
- With usage and quality reporting
One question, start to finish.
Five steps, from question to answer. Also try a question you shouldn't be able to ask.
Permission first, answer second.
it answers in Teams or WhatsApp, with no SQL and no new reports.
your company's permissions are checked before any query: what's off-limits is explained, not shown.
several companies or departments on the same platform, without one seeing another's data.
it's deployed as code in your company's cloud: your data doesn't go through ours.
Every release is tested before it ships: 29 of 30 on accuracy with real questions, and 30 of 30 on permissions with different users.
What people ask us.
Does my data leave my cloud?
It never leaves a cloud you own. The agent is deployed as code in your account and queries your lakehouse where it already lives. Today it runs on AWS: if your data lives in another cloud, we replicate what's needed to a secure account that's also yours, or we build it on your main cloud.
What happens if someone asks about something they're not allowed to see?
Before any query runs, their table- and column-level permissions are checked. If they don't have access, the agent explains why it can't answer.
Do I need a lakehouse first?
It helps to have one organized in layers. If you don't, our Data Engineering practice helps you build it the right way: layered, governed, and with its semantic layer.
How do I get it?
It's part of the monthly service of your pod, the Cloudih team that works with you. There's no separate license.
With the tools you already use.
Start with an assessment.
Within weeks, you know where the value is. From month one, you see it.
Book your assessment Explore our services in more detail →