Data Strategy Assessment
We read your estate as it runs, typically in two weeks. You keep a costed modernization map with its TCO delta.
Governance decided
before the next workload lands.
Decide who owns each domain, who may see what and how lineage is traced, with every rule mapped to where Fabric or Databricks enforces it.
Owners, named
One accountable owner for every domain.
Access, enforced
Each rule tied to the platform feature that enforces it.
Lineage, baselined
What the platform traces today, and where the gaps are.
Part of the Blueprint phase, typically two to three weeks · Fixed price · The model is yours
Set the rule first,
and every new table inherits it.
Usually the platform ships first and governance waits for next quarter. By then access was granted one ticket at a time, copies sit in several workspaces with no owner, and lineage is pieced together by hand.
Decided before the next workload
A grant on a schema covers every table created there later, so the rule has to come first.
Owned by who answers for it
Domains follow the people accountable for the data, not the org chart.
Ready for agents
Agents act with the access they are given, so this is the model your first agent works under.
This is the right engagement when
- A Fabric or Databricks estate is in design, and no workload has landed.
- A Unity Catalog upgrade or Fabric move is funded, without an access model.
- Access was granted person by person, and no list shows who holds what.
- Agents are on the roadmap, with no access model for them to work under.
Every rule,
one enforcement point.
Five steps inside the Blueprint phase take you from who owns the data to the backlog that puts the rules in.
We inventory workspaces, catalogs and the grants held today. With your leadership we settle the domains and one owner for each.
You receive
Domain and ownership map
Every domain with its one accountable owner.
What this is not.
This work ends in a governance model. Putting it in, and checking each agent request, happens elsewhere.
Putting the rules in during a move?
That is ModernizationChecking what an agent may do on each request?
That is Trust LayerDefining metrics inside semantic models?
That is Analytics & BI
The story this service stands on.
CenteredCare
Healthcare · Senior living
- To onboard a new community, down from 3 weeks
- <24h
- Fewer ingestion errors
- 60%
How CenteredCare Onboards a New Community in Under 24 Hours
CenteredCare needed every senior-living community's clinical data on one platform, with protected health information kept inside a governed path. We gave access control and lineage explicit places in the platform from the start: Azure Key Vault, role-based access control, Microsoft Entra ID and Microsoft Purview. A new community now onboards in under 24 hours instead of 3 weeks, with 60% fewer ingestion errors.
Questions buyers ask.
We already own Purview or Unity Catalog. Why pay for a design?
Owning a catalog leaves the operating decisions open: who owns a domain, which groups may use its data, how access is granted and what lineage the team needs. This engagement settles those decisions and maps them to the platform. Platform permissions enforce access; the catalog and lineage design help the team understand the data and inspect how it is used.
Do we have to choose between Fabric and Databricks first?
No. Domains, owners and access tiers can be decided before the platform and become requirements for Technology Selection. Once the platform is chosen, each rule is mapped to the controls it provides. The design also records the lineage the platform captures and the gaps that need instrumentation, so the implementation starts from explicit requirements and known limitations.
Our estate is already live and was never governed. Is it too late?
No, but the order changes. The design starts from the grants, workspaces and catalogs you have today and sequences the path from today's access to the target model. GovKit moves a Hive or Glue metastore to Unity Catalog carrying users, ACLs and jobs, with testing. On a live estate that is not moving, the uplift runs as the same fixed-scope engagement. Lineage is the part you cannot backfill: Unity Catalog records it only from queries run after a table comes under its metastore.
Does this cover our AI agents?
It covers the access an agent works under: which groups and service principals can see which domains, tiers and columns. A Fabric data agent, for example, answers with the asking user's permissions, row and column security included, so the access model decides what it can reach. The checks that run on each agent request, identity per action, tool scopes, content safety and human sign-off, belong to the Trust Layer in Data & AI, which takes this design as given and never redesigns it.
How long does it take, and what does it cost?
We read your estate through the Assess phase's scan, which typically runs two weeks; the governance model is one deliverable of the Blueprint phase that follows, which typically runs two to three weeks. Each phase is fixed price, fixed scope, priced up front, and the model is yours whether or not you continue. The uplift that carries it out is a separate fixed-scope engagement per workspace estate.
More in Strategy & Advisory.
Natural next step
Natural next step: Technology Selection, Strategy & Advisory
When the governance model is signed, it crosses to Technology Selection, and the question changes from what the rules are to which platform enforces them best on your workloads.
Technology Selection
Fabric, Databricks or the platform you run today, settled on the few workloads that carry the decision, before any migration is funded.
Data & AI Strategy
Every candidate use case gets a written verdict. The survivors become a costed roadmap and delivery backlog.
AI Readiness
One agent is measured before it is built: Butter's free verdict, then typically two weeks on its data, controls and evaluation baseline.
Bring the access
you have already granted.
Bring your platform diagram and any record of who holds access today to a 30-minute working session.