The most expensive AI mistake I see is a platform bought off a demo and handed to a team that never asked for it. The second most expensive is doing nothing while a few people quietly build things in tools nobody approved.
The assessment sits between those. I talk to every department, look at how work moves between them, and come back with a prioritized list: what is worth automating now, what needs a tool, what needs a process fix first, and what to leave alone. Most of what I find is not AI work at all, and I will say so.
Who this is for
- Leadership has asked for an AI plan and you do not yet know where the time and energy should go.
- Vendors are pitching and you want an independent read before signing anything.
- Your departments run on different systems and nobody has looked at how the work connects.
Before any of this starts
Every engagement starts the same way, whichever one you pick. It takes a few conversations, on purpose: you are choosing someone your teams will work with directly, and the scope comes out of those conversations.
- 1
Intro call
You
Thirty minutes to hear where your company is with AI, what leadership is asking you for, and what you are trying to get done. It is enough to tell whether a longer conversation is worth having, and that is all it needs to do.
- 2
Discovery sessions
You first, then your leadership team
Usually two or three, each about an hour. The first is with you: your issues, priorities, and concerns, properly, often with some hands-on work so you can see how I operate. Then a level-set with the people who will have to approve this, on a call or in person: my approach, what I expect from your side, and what the engagement would look like. This is where the scoping starts.
- 3
Proposal and statement of work
You and leadership
Based on what came out of discovery, I send a short proposal: what we would do, in what order, and what it costs. Once everyone is comfortable with it, it becomes a signed statement of work.
- 4
The work starts
Everyone
Sessions get scheduled around your priorities and your team's availability. Everything I produce lives in your systems, and you own it whether or not we work together again.
How it runs
Once the scope is agreed, this is the work.
- 1
Agree what success looks like
Leadership, often in the same meeting where we agreed to work together
An hour to write down what a good outcome means here: hours back, faster cycle times, fewer manual handoffs, less risk. Without that, a roadmap has nothing to be scored against, and the first thing a good roadmap does is tell you what not to do.
- 2
Systems and data inventory
IT, and whoever knows where the data lives
A short pass over what you run, what has an API or an export, and where the data is clean enough to use. Half of what people want to automate turns out to depend on a spreadsheet nobody trusts. Knowing that early keeps the recommendations honest.
- 3
Team lead conversations
Each department lead, and a few people from their team if they want to join
A short call per department to hear where they feel the roadblocks are, what they have already tried, and how ready their people are. Leads see the bottlenecks. The people doing the work see the workarounds. I want both.
- 4
Research
Me
Between the lead calls and the workshops I go and do the work: map the processes I heard about, look at the tools already in the building and what they can do that nobody has switched on, and check what the vendors circling you can and cannot do. This is where the assessment earns its independence.
- 5
Department workshops
Each full department
One to four hours per group, depending on the size of the department and how much of its week is process. We walk through how the work runs, where it stalls, where people copy things between systems by hand, and what they would fix first if they could. I ask a lot of questions and take a lot of notes. Nobody has to prepare anything.
- 6
Follow-ups
Groups or individuals, as needed
Some things only become clear afterward. A process that looked simple in the workshop has an exception path someone mentions in the hallway. I run short follow-ups per group or per person to close those gaps before I write anything down.
- 7
The assessment and the readout
You and leadership
A written assessment and a prioritized roadmap. Quick wins you can do this month. Larger automations worth scoping. Where data needs cleaning up before anything else. Which tools or integrations to look at, and which to ignore. What not to touch, and why. A draft usage policy alongside it, so the roadmap comes with rules. Then a readout with leadership where we go through it and I answer whatever comes up. You own the document whether or not we work together again.
What you get
- An agreed definition of what success looks like, written down before any work starts
- A systems and data inventory: what you run, what connects, what is clean
- Notes from every department on how the work runs and where it stalls
- A prioritized roadmap: quick wins, larger builds, data cleanup, and what to leave alone
- A draft usage policy, so the roadmap comes with rules
- An independent read on the tools and vendors in front of you
- A readout with leadership, and a document you own afterward
What it takes from your team
About an hour from each department lead, one to four hours from each department as a group, an hour from leadership at the start and again at the readout. IT spends an hour or two on the inventory. Nobody prepares slides.
What it does not include
Building what the roadmap recommends. The assessment tells you what to do and in what order. Doing it is either the enablement program, fractional support, or a scoped build, and the roadmap says which fits each item.
Is it a fit?
Thirty minutes, no pitch deck. Tell me where your company is with AI and what leadership is asking you for, and I'll tell you honestly whether this is the right shape, or whether one of the other two fits better.