Find out where AI actually fits in your environment
Thirty minutes with the director who runs the practice. You leave knowing which of your data and workloads are ready, what has to change first, and which next step is worth paying for.
- Enterprise software since 1994
- Microsoft Cloud Solution Provider
- CISSP and CISA certified security leadership
- Fortune 500 clients across North America
Most AI programs stall before the first build
Organizations have run the experiments. A Copilot pilot, a proof of concept on a single document set, a model somebody built in a notebook. The harder question comes next, and it isn’t a modeling question. It’s where the data lives, who is allowed to see it, and which workload is worth building first.
That question gets answered badly in two directions. Buying a discovery engagement to answer it is expensive and slow. Skipping it means the first build lands on the wrong use case, or on data nobody has cleared for it.
Thirty minutes, one honest read
The assessment is a thirty-minute conversation with Eddie Hudson, who runs Winmill’s Emerging Technology practice. You describe your data, your environment, and what you want AI to change. You leave with a clear read on what’s ready, what has to change first, and whether a next step exists that’s worth paying for.
It needs no access to your systems, it costs nothing, and it isn’t a gate to a proposal. If nothing we sell fits, that’s a legitimate outcome.
Six things we look at
Where your data lives
The systems holding the data an AI use case would read, and what condition it’s in. Fabric, SQL, SharePoint, line of business applications, or all four.
Identity and access
Who is permitted to see what today, and whether an AI interface would respect those rules or quietly route around them.
The use case and the measure
What the work is supposed to change, and how you’d know it worked. A use case with no measure is the most common reason a pilot ends without a decision.
Platform fit
Fabric, Azure AI Foundry, Azure Machine Learning, and the container platform questions underneath them. This is where an existing AKS estate or a Container Apps decision gets factored in rather than ignored.
Security and compliance boundaries
What your obligations are, what your tenancy allows, and what that rules out. Every claim we make about scope is written down rather than implied.
What has to change first
The prerequisites. Usually data access, sometimes identity, occasionally a platform decision that has been deferred.
Three outputs, one of them uncomfortable
A clear read on your environment
Not a proposal. Eddie tells you what’s ready, what isn’t, and why, in the call itself.
A named next step
If there is one, it’s an offer that exists and has a shape you can approve: a Fabric Fit Check, a fixed price pilot against one data source, or taking over models already running in production.
An honest no when it’s a no
Sometimes the answer is that the data work has to come first and AI is a year away. We say that in the assessment rather than after you’ve paid for a phase.
How this stays useful
It isn’t a paid discovery phase
The assessment costs nothing and it isn’t a gate to a proposal. If nothing we sell fits, that’s a legitimate outcome.
Nothing leaves your environment
The assessment is a conversation and a review. It needs no access to your data and no connection into your tenancy.
Security is in-house
Penetration testing and code audits are Winmill practices. When we assess whether an AI interface respects your access rules, that reading comes from the team that tests systems for a living.
Where it usually goes next
Your data is in reasonable shape
The Fabric Fit Check is free and takes under two minutes, and it tells you whether Fabric fits before anything is scoped. See Microsoft Fabric and Foundry.
You want one use case proven
AI on Your Data starts with one library and one interface, and the pilot is step two of four. See AI on Your Data.
You already have models running
If models are in production and nobody owns them, that’s an operations problem rather than a build. See Machine Learning Operations.
Frequently asked questions
Is the assessment really free?
Yes. It’s thirty minutes with the director of the practice, and there is no charge and no obligation. Paid work starts only if you approve a scoped next step.
Who should be in the room?
Whoever owns the data and whoever owns the outcome. That’s often two people. A platform lead is useful when the question involves Azure architecture.
What if we’re not on Azure yet?
Say so at the start. The assessment is still useful, and the honest answer may be that the migration question comes before the AI question.
Do you need access to our systems?
No. Nothing connects to your environment. The assessment works from what your team describes and what you choose to show.
What happens if you find we’re not ready?
We tell you, and we tell you what would have to change. That outcome ends more of these assessments than a signed pilot does, and it’s the reason the assessment is worth taking.
Thirty minutes, and an honest read
Book the assessment with Eddie Hudson, who runs Winmill’s Emerging Technology practice and takes these calls himself.