Enter your search text in the box above

Select a result to preview

The Fabric Effect at CDW: Built by the Business, Supercharged by IT

#fabcon2026 #session #learning #microsoft #fabcon2026-session

Description

See how CDW brought business and IT together to build something bigger. Christopher shows how he used Fabric to turn scattered data into a single operational portal that runs the business. Julie adds technical insights and reveals how partnership unlocked even more value. A fast, energetic look at the power user revolution inside Fabric.

Key Takeaways

My Notes

Action Items

Slides

The Fabric
Effect at CDW
Built by the Business, Supercharged by IT
Christopher Marcolis · Head of Presales, Data & AI
Mwazanji Sakala · Sr Solution Engineer. Data & AI
Y OU R HOS TS
Two Voices. One Story.
CHRISTOPHER MARCOLIS
MWAZANJI SAKALA
Head of Presales
CDW Data & AI Practice
Sr. Solution Engineer
CDW Data & AI Practice
"My business couldn't wait ## months
on the IT backlog. So, I built it myself.
With duct tape, hairspray, and unicorn farts."
"What Christopher just called 'duct tape,
hairspray, and unicorn farts.' — I'll translate that
into actual architecture for you."
The Executive Narrator
The Technical Translator
The Analytics Problem
(Every Business Leader Knows)
Data is Everywhere
IT Can't Keep Up
The Backlog is Real
CRM. ERP. HCM. Marketing.
Finance. Every team built their own
island.
Innovation is accelerating faster
than enterprise architecture review
cycles.
##-month waits. Competing
priorities. Business needs can't
pause for the queue.
"I had data in five systems. I needed one answer. IT was going to get to me in Q3... of next year." — Christopher
SHA DO W IT CO NFE SSIO N
I Couldn't Wait. So, I Built It!
I run a data-driven presales operation — pipeline,
partnerships, seller performance
My data lives in Salesforce, Excel, Snowflake, SharePoint,
and 2 homegrown tools

mo

Source systems
to integrate
IT backlog
wait time
100%
$15M+
Held together
by favors & duct tape
Pipeline this
thing now tracks
I needed a single view of my business. IT had me in the
backlog
So, I called in favors. Used consulting hours. Wrote some
glue code. Built it myself.
Used a semantic layer to stitch it all together. It worked.
Mostly.
M WAZ ANJ I ' S T A KE
What Christopher built was a Power BI semantic model pulling from shared OneDrive + Salesforce exports. It worked — but every
connector update was a 2am fire drill.
TE CHN IC A L RE AL IT Y C HEC K
What 'Glue & Hairspray' Actually Means
P RE SE NT A T IO N
SE MA NT I C
I NT E GRA T I O N
SOU RCE
Power BI Reports + Manual Excel
Power BI Dataset (the 'semantic glue')
The Real Problems
Manual refresh = stale data at the worst
moment
Any connector update breaks the chain
Power Query + Manual Refresh Schedules + Shared
OneDrive CSVs
No lineage — nobody knows what broke
Salesforce · Snowflake · SharePoint · Excel · Homegrown
App
IT won't touch it. Can't support it.
Single person dependency to fix it
CH RIS TOPH ER’S TAK E
JOINat
THESalesforce had a bad day.
I didn't know what half of this was called. I just knew it broke every
time someone
It Worked. Mostly.
The 2am Connector Break
Salesforce updated their API.
My entire pipeline went dark.
Monday morning QBR had no
data.
The Data Drift
Single Thread of Failure
Nobody noticed for 3 weeks
that one of my Excel files got
renamed. Metrics were
wrong. Pipeline calls were
wrong.
I was the only one who could
fix it. That's not a data
problem. That's a business
risk.
The system I built to run my business was becoming the thing I had to run
instead of my business.
THE PI VO T
I Went Back to IT.
But this time, I came differently.
HO W I U SE D T O AS K
HO W I AS KE D T HI S T IME
"Build me a thing"
"I already built a working prototype"
"I need it by next sprint"
"Here's what I need governed and supported"
"It should look exactly like this"
vs
"I'll compromise on design if you'll sponsor it"
"IT is blocking me"
"I want to be in Fabric — your stack, your rules"
"Just give me access and I'll handle it"
"Let's build something together that IT can own"
IT Said Yes. With Conditions.
Use the platform as designed
Compromise on the design
No more side-loading. Dataflow Gen2, OneLake, and
approved connectors only.
My opinionated architecture had to bend to Fabric's
patterns. That's the deal.
IT sponsors, not just tolerates
Make it reusable
In return for compliance, IT becomes a real partner —
monitoring, support, scale.
This can't just serve presales. The architecture has to be
a template the business can use.
This moment — an IT team willing to sponsor a business-led build — is rarer than it should be.
THE PL A TF O RM SHIF T
Why Fabric Changed the Equation
Before Fabric
With Fabric
Business builds in shadow IT
One platform. One tenant. One lake.
IT governs in a separate stack
Business user and IT live in the same house
The two never meet cleanly
IT can govern what I build
Semantic layer = the war zone
I can move fast inside guardrails
M WAZ ANJ I ' S T A KE
Fabric isn't just a tool. It's the first platform where the power user and the platform engineer can
coexist.
Fabric's
unified compute and OneLake storage means the semantic model I build as a citizen developer lives in the same governed
environment as enterprise pipelines. No export-to-CSV handoff required.
ARCHIT ECTUR E: WHAT W E BUI LT T O GETHE R
The New Stack — Rebuilt Inside Fabric
CO N SU M E
Power BI Direct Lake Semantic Model · Direct Lake Reports · Teams/SharePoint Embed
S E RV E
Fabric Warehouse · Dataverse Tables for Transanalytics · Lakehouse SQL Endpoint
TR ANS FO R M
Dataflow Gen2 · Fabric Notebooks · Data Pipeline Orchestration
STORE
OneLake · Delta Parquet Tables · Lakehouse (Bronze / Silver / Gold)
I N G E ST
Dataflow Gen2 Connectors · Salesforce · Snowflake · SharePoint · Dynamics 365
CH RIS TOPH ER’S TAK E
↑ This is what 'glue, hairspray and unicorn farts'
looks like when
it grows
up. ↑
THE CO NN EC TO RS THA T UNL OC KED EV ER YT HING
Three Capabilities. One Breakthrough.
Dataflow Gen2
OneLake Storage
Dataverse Tables
EXEC
EXEC
EXEC
Connects to everything I used to
manually export. Certified connectors,
scheduled refresh, no code required.
One place to put everything. IT governs
it. I use it. Nobody argues about where
the data lives.
My operational data — live. Not a
report. An actual table that writes back
to the business system.
M WA Z A N JI 'S T AK E
M WA Z A N JI 'S T AK E
M WA Z A N JI 'S T AK E
Power Query-based ETL with Output
Destinations to Lakehouse or Warehouse.
Supports 150+ connectors including
Salesforce, Dynamics, Snowflake. Staging on
OneLake by default.
Delta Parquet on ADLS Gen2, hierarchical
namespace. Fabric Shortcuts enable zerocopy data sharing across workspaces. No
more data silos.
Transactional + analytical access via
DirectLake. Power Apps, Power Automate,
and fabric pipelines all talk to the same
table. Zero ETL between operational and
analytical.
One Portal.
One Truth.
Pipeline Visibility
Partner Scorecards
100s opportunities. $15M.
Live. One screen.
Snowflake, Microsoft,
Databricks — all tracked, all
current.
Seller Performance
Presales Capacity
Attach rates, opportunity
quality, coverage by vertical.
Who's working what, where
we're overloaded, where we
have room.
"This is the tool I run my business from.
Every pipeline call.
Every partner review.
Every leadership ask
— answered in seconds."
M WAZ ANJ I ' S T A KE
Under the hood: DirectLake semantic model on Lakehouse Gold layer + Dataverse virtual tables for write-back. Power BI Embedded in a
Teams app. IT monitors the pipeline. BU owns the report layer.
THE IMP A CT
Before & After: What Actually Changed
D I ME N S I O N
BE FORE — Gl ue & Ha i r spr a y
AFT ER — Fa br ic + I T Pa r t ner s hip
Data Freshness
Automated & Manual refresh yo-yo
✓ Scheduled Dataflow Gen2 · near real-time
Connector Breaks
2am fire drills when APIs change
✓ Certified connectors · IT monitors
Time to Insight
Build a new report: 2–3 weeks
✓ New report on governed model: hours
IT Relationship
IT won't touch it
✓ IT sponsors and co-develops
Reliability
Single person can fix it (me)
✓ Documented, governed, transferable
Scale
Only works for presales
✓ Architecture template for any BU
M WAZ ANJ I ' S T A KE
The technical unlock: moving from 'dataset in a workspace nobody governs' to 'Lakehouse + Warehouse in an IT-sponsored capacity'
changes everything about reliability, monitoring, and cost transparency.
WHAT WE L EARNE D
How I Learned to Behave.
(And why IT became my best partner.)
Show up with a working prototype, not a requirements doc. IT respects builders.
Compromise on architecture. If IT's going to own it, it must be built on their platform.
Shadow IT isn't the enemy — it's R&D. Treat it that way, then bring it home.
IT sponsorship isn't just about support. It means the work lives past you.
The power user era isn't over. It just needs a platform that can hold it.
M WAZ ANJ I ' S T A KE
The fastest path to a governed, production-grade data platform is a business user who already built
the prototype — we just swap the
duct tape for Dataflow Gen2 and put IT's name on the capacity.
MWAZ ANJI 'S T AKE
The Power User Revolution
is Already Inside Your Organization.
Your best data builders
…are in the business, not in
IT. They just need the right
platform.
Shadow IT is prototyping
…the most relevant, highvalue use cases in your
organization.
Fabric is the bridge
…that lets business agility
and IT governance live in the
same house.
Stop fighting your power users. Start building a platform that deserves them.
The Takeaway
IT can be
your best
co-developer.
WH AT 'S NE XT
Copilot pages over the portal
Fabric Real-Time Hub for pipeline events
Cross-BU reuse of the architecture
Fabric Agent Skills for auto-insights
If you learn to behave.
Find us at the CDW booth
#FabricEffectCDW
Christopher.Marcolis@cdw.com
Mwazanji.Sakala@cdw.com
Thank you. Questions? Let's talk.
How was
the session?
Complete Session Surveys in
for your chance to WIN
PRIZES!
Sound off.
The mic is all yours.
Influence the product roadmap.
Join the Fabric User Panel
Join the SQL User Panel
Share your feedback directly with our
Fabric product group and researchers.
Influence our SQL roadmap and ensure
it meets your real-life needs
https://aka.ms/JoinFabricUserPanel
https://aka.ms/JoinSQLUserPanel