AI assistant — not human

Ranked by delivery depth: F-SKU capacity planning, OneLake + DirectLake architecture, Lakehouse migration, Real-Time Intelligence, Purview governance.
Last updated July 9, 2026 by Errin O'Connor, Founder & Chief AI Architect, EPC Group
Six firms with substantial Microsoft Fabric delivery depth: EPC Group (Houston, 29 years Microsoft data platform, all six Solutions Partner designations), Netwoven, Pragmatic Works, Slalom Analytics, BlueGranite, Concurrency. Seven Fabric-specific capabilities to evaluate: F-SKU capacity planning, OneLake architecture, DirectLake expertise, Data Factory + Copy Job, Real-Time Intelligence + Eventhouse, Fabric CI/CD, and governance. Cost benchmarks: departmental pilot $65K-$185K through enterprise platform $650K-$1.8M. Seven pitfalls to avoid: under-sizing F-SKU, DirectLake conversion issues, skipped shortcut strategy, reversed Warehouse/Lakehouse decision, deferred Purview, no CI/CD, no cost mgmt dashboards.
Six firms with substantial Microsoft Fabric delivery depth as of 2026: (1) EPC Group — Houston-based boutique specialist, all six Solutions Partner designations (Data & AI + Infrastructure + Digital & App Innovation + Business Applications + Modern Work + Security), 29 years Microsoft data platform delivery from SQL Server through Power BI Premium P-SKUs to Fabric F-SKUs. (2) Netwoven — California-based, cross-M365 depth with strong Fabric practice. (3) Pragmatic Works — Florida-based, Microsoft data platform training + implementation, historical SQL Server / SSIS depth. (4) Slalom Analytics — regional bench delivery, Microsoft + multi-vendor. (5) BlueGranite — mid-market Microsoft data platform boutique. (6) Concurrency — Wisconsin-based Fabric + Azure Data platform focus. Big Four Fabric practices (Deloitte Data + AI, PwC Cloud + Digital, EY Data + Analytics) also deliver — typically for board-level programs with brand credibility priority. Selection factors: implementation count, F-SKU sizing depth, migration methodology, senior consultant continuity, regulated-industry depth.
Seven Fabric-specific capabilities that separate deep firms from resellers: (1) F-SKU capacity planning + rightsizing — F2 through F2048 sizing methodology, workload distribution modeling, cost projection accuracy. (2) OneLake architecture — Lakehouse vs Warehouse selection, shortcut strategy, cross-workspace data mesh design. (3) DirectLake mode expertise — dataset conversion from Import + DirectQuery to DirectLake, performance tuning, refresh strategy. (4) Fabric Data Factory + Copy Job — pipeline design, incremental load patterns, error handling, monitoring. (5) Real-Time Intelligence + Eventhouse (formerly Real-Time Analytics + KQL Database) — stream ingestion, KQL query patterns, event-driven dashboards. (6) Fabric CI/CD — deployment pipelines + Git integration + workspace governance. (7) Fabric governance — Purview integration + sensitivity labels + Domains + workspace routing. Ask for evidence + case studies for each — Fabric is a broad platform and firms specialize.
Seven engagement sizes: (1) Fabric Readiness Assessment (fixed-fee, 3-5 weeks, produces target-state F-SKU + architecture + roadmap) — $25K-$65K. (2) Departmental Fabric pilot (single workload, F2-F8 capacity, 3-5 dashboards) — $65K-$185K, 8-14 weeks. (3) Mid-market Fabric implementation (F16-F64, 3-5 workloads, cross-department semantic model) — $225K-$650K, 16-32 weeks. (4) Enterprise Fabric platform (F64-F512, medallion architecture, CoE governance, DevOps + CI/CD) — $650K-$1.8M, 6-14 months. (5) Power BI Premium to Fabric F-SKU migration (dataset conversion, DirectLake enablement, cost optimization) — $150K-$550K, 12-26 weeks. (6) Legacy warehouse to Fabric Lakehouse migration (Synapse Dedicated SQL Pool, Snowflake, Redshift, on-prem SQL warehouse) — $350K-$2.5M, 6-18 months. (7) Fabric managed service retainer — $15K-$85K/month. Fixed-fee model — no T&M sprawl.
Seven common pitfalls: (1) F-SKU under-sized at pilot — pilot on F2 that can't scale, need to re-architect on F16+. (2) Import datasets migrated to DirectLake without validation — DirectLake requires Delta tables + specific size constraints; not all workloads convert cleanly. (3) OneLake shortcut strategy skipped — data duplication across workspaces + capacity waste when shortcuts would consolidate. (4) Warehouse vs Lakehouse decision reversed after implementation — costly re-architecture; decision should happen at architecture design phase not build. (5) Purview integration deferred — governance retrofitted later is 3× more expensive than governance designed in from Day 1. (6) CI/CD not implemented — manual deployment through UI causes production incidents + no rollback capability. (7) Cost management dashboards not built — capacity units burn without visibility, budget overruns discovered at month-end. EPC Group's Fabric Readiness Assessment methodology explicitly addresses each.
Six EPC Group Fabric credentials: (1) All six Microsoft Solutions Partner designations — Data & AI is required for Fabric; the additional 5 signal cross-workload depth. (2) 29 years of Microsoft data platform delivery from SQL Server + SSIS + SSAS + SSRS through Power BI Premium P-SKUs to current Fabric F-SKUs. (3) Errin O'Connor — Chief AI Architect + Microsoft Press bestselling author of "Microsoft Power BI Dashboards Step by Step" (Fabric-era edition) + pre-release program participant for Power View / Project Crescent (the codename that became Power BI, which is now integrated into Fabric). (4) Individual senior consultant certifications — DP-600 Fabric Analytics Engineer Associate + DP-500 Enterprise Data Analyst + PL-300 Power BI Data Analyst. (5) 1,500+ Power BI implementations underpin the Fabric practice — Fabric consultants without Power BI depth deliver worse outcomes. (6) G2 Leader in Business Intelligence Consulting (6 consecutive quarters Fall 2024 through Summer 2026).
Six-phase methodology for evaluating + executing a Fabric program: (1) 30-min qualification call — mutual fit assessment. (2) Phase 1 Fabric Readiness Assessment (fixed-fee $25K-$65K, 3-5 weeks) — current-state Power BI + warehouse assessment, F-SKU sizing model, target-state architecture (OneLake + Lakehouse + Warehouse mix), migration wave plan, fixed-fee full-program proposal. (3) Phase 2 Fabric platform foundation (F-SKU provisioning, Purview integration, DevOps + CI/CD pipeline, governance workspace routing) — fixed-fee, 4-8 weeks. (4) Phase 3 workload migration in waves — pilot → validation → production rollout. (5) Phase 4 optimization + rightsizing — F-SKU consumption analysis, capacity redistribution, workload isolation. (6) Phase 5 sustainment retainer — ongoing platform ops, new workload onboarding, capacity monitoring. Deliverables: F-SKU sizing model, target-state architecture, migration runbook, DAX + KQL measure library, RLS rules, sensitivity labels, CI/CD pipelines.
Big Four Data + AI practices (Deloitte, PwC, EY, KPMG) deliver Fabric at scale for board-level programs where brand credibility matters. EPC Group's advantages in the same market: (1) 30-40% lower blended rates. (2) Senior consultant leadership throughout — Big Four programs often mix partner + senior manager + senior consultant + offshore associate; EPC delivers senior-consultant-only. (3) Faster mobilization — 2 weeks vs Big Four's 4-6. (4) Microsoft-only depth — no context-switching cost with multi-vendor practice areas. (5) Direct access to firm founder for architecture decisions. (6) Deeper regulated-industry vertical proof (healthcare / financial services / defense) than most Big Four Fabric practices. Big Four wins when: multi-year 3-5 year transformation with $2M+/year commitment, global multi-region delivery, incumbent Big Four relationship, board reporting where brand credibility is decision-critical.
30-min qualification + Fabric Readiness Assessment ($25K-$65K fixed-fee, 3-5 weeks). Call (888) 381-9725.
Monday-Friday, 8 AM - 7 PM CT
We respond to all inquiries within one business day