Microsoft Fabric, data, and AI
Most AI projects stall on the data underneath them. We design, build, and govern Microsoft Fabric platforms so the numbers are trusted, the cost is understood, and Copilot and data agents have something solid to stand on.
See what we doWhat we do
Seven practices, each a defined body of work with named deliverables. Most engagements use two or three. Open any card for the detail.
Microsoft Fabric
Microsoft Fabric
Workspace topology, capacity model, and medallion design decided before anything is built on top of them.
Microsoft Fabric
Microsoft Fabric
Synapse, Tableau, and legacy SQL estates moved onto Fabric with the business logic intact and the numbers reconciled.
Microsoft Fabric
Microsoft Fabric
Where every capacity unit goes, who should pay for it, and what the SKU should actually be.
Data
Data
Pipelines, lakehouses, and real-time ingestion built to be operated by your team rather than babysat by ours.
Data
Data
Star schemas, DAX that holds up under concurrency, and a semantic layer the business trusts.
Data
Data
Purview cataloging, sensitivity labeling, and access control built into the platform rather than bolted on afterwards.
AI
AI
Copilot in Fabric and Power BI, Fabric data agents, and the model and governance work that decides whether their answers can be trusted.
Our product
Lakewarden
Access, capacity, refreshes, connections, and tenant change — a whole Fabric estate in one console, signed in with your own Microsoft identity.
Built for the practices on this page. Each module covers the ongoing operating side of one of them.
Governance and security
Access Review
Every principal, every workspace, every role — guests flagged, effective access resolved through group membership, and role changes previewed before a write.
ExploreCapacity and cost optimization
Capacity and CU usage
Per-item CU attribution, hour-by-weekday load, spike forensics, and SKU what-if — so a capacity conversation runs on evidence instead of instinct.
ExploreData engineering
Refresh history
The last fifty runs of every model with Fabric's own error text and a diagnosed cause, so a failing refresh is a known problem rather than an investigation.
ExploreMicrosoft Entra verified publisher.
Lakewarden is an independent product. It is not affiliated with or endorsed by Microsoft.
Reference architecture
Source systems through to the semantic layer and the agents that read it, with governance, security, operations, and delivery treated as part of the architecture.
Security
Entra identity, workspace roles, row and object level security, private networking, and tenant policy baseline.
Operations
Orchestration, structured run logging, alerting, capacity monitoring, chargeback, and quarterly right-sizing.
Delivery
Git integration, deployment pipelines, environment parameterization, and reviewable promotion to production.
03OneLake
A single logical lake in open Delta Parquet. Every Fabric engine reads and writes the same tables, which removes the copy-per-tool pattern that makes reconciliation impossible and storage costs compound.
How we work
The sequence is deliberate. Each step is only as good as the one before it, and the last one is the point of the other four: your team running the platform without us.
Start with the free review01
A free thirty-minute conversation about your current estate, what you are trying to reach, and whether Fabric is genuinely the right answer. No deck, no pitch.
02
Workspace and domain topology, medallion structure, security model, and capacity plan get decided while they are still cheap to change. Documented and reviewed before anything is built on top of them.
03
Delivery in two-week increments against a defined backlog. Pipelines are idempotent, models are version controlled, and every increment ends with something your team can see running.
04
Logging, alerting, access review, cost controls, and a validated recovery path. Migrations are reconciled against the source system before anything is retired.
05
Knowledge transfer is scheduled work, not a document dropped on the last day. The engagement ends with your team operating the platform without us.
Free consultation
We look at the architecture you have today, what you are trying to reach, and whether Microsoft Fabric is the right platform for it. If it is not, we will say so.
What we cover
Based in
Minneapolis, Minnesota
Working with teams across the United States
