# Lakehouse Launchpad

> The Launchpad productises how we stand Databricks up for real customers: Unity Catalog before anything else, ingestion that holds, layers that mean something, dashboards people open, and a Genie space with a curated semantic layer that answers real questions.

For: Teams modernising off a legacy warehouse, or stalled after a pilot
Length: Six weeks
Canonical: https://www.techfabric.com/databricks/launchpad

---

A governed lakehouse your executives can question in plain English

## What you get

- A workspace foundation with an identity and environment strategy you can defend
- Unity Catalog structure and the governance patterns that go with it
- Ingestion paths through Lakeflow or Jobs, whichever suits the source
- Bronze, silver and gold layers for the domains you named at the start
- AI/BI dashboards for the people who asked for them
- A working Genie space with real semantic definitions
- Runbooks, and your team able to use them

## Questions

### Is six weeks realistic, or is it a sales number?

It is the programme length, and it holds because the scope is fixed at the start. What moves the date is access: if credentials and source system owners take three weeks to appear, the clock reflects that. We say so during scoping.

### What if our data is messier than we admitted?

It usually is, and the two-to-three week discovery before the build exists to find that out while it is still cheap. If discovery shows six weeks is not honest for your estate, we say so and rescope before anything starts.

### Do you leave, or do we get stuck with you?

The runbooks and the enablement are deliverables, and the point of them is that your team can run this without us. Plenty of clients keep us on for the next programme, but that should be a decision rather than a dependency.

### We already have a lakehouse. Is this still relevant?

Then you probably want the Health Check instead, which assesses what is there before anyone proposes rebuilding it. The Launchpad is for standing something up, not for replacing a platform that works.