Your IT Department Is Not the AI Strategy

Theme 3, Article 7 — 2weekAI Blog

Let me tell you about a Tuesday in 1995.

I was an Assistant Vice President at MBNA America. One of the most profitable banks in the country. And every manager, every officer — regardless of title — spent four hours a month in the call center. Taking calls. Reading customer letters. Sitting next to the people doing the work every single day.

Not observing. Doing.

So there I am, headset on, customer call just disconnected. And I notice something wrong. A field level edit in the workflow is broken. Something the next customer is going to hit. Something that's going to create a problem that shouldn't exist.

I didn't file a ticket. I didn't send an email to my manager. I didn't schedule a meeting to discuss the issue and evaluate the impact and assemble a working group to assess options.

I picked up the phone.

"Hey Tony. The field level edit is broke. Can you fix it?"

Five minutes later.

"Dave, try it on the next call."

Fixed. Done. Next customer.

What Happened to Tony

Tony still exists. He's just buried under a change control process, a ticket queue, a security review, a sprint backlog, and three layers of approval that didn't exist in 1995.

The problem still gets found. It just takes six weeks to fix now instead of five minutes.

And somehow we've decided that's the cost of doing business at scale.

It isn't. It's the cost of building organizations where the people who find problems don't have direct access to the people who can fix them. Where the information gap between the front line and the decision maker is so wide that by the time the problem arrives with the right person it's been sanitized, prioritized, deprioritized, reprioritized, and scheduled for next quarter.

I've spent 28 years watching this happen across organizations of every size. Billions of dollars in projects delivered. The problem isn't the technology. It's never been the technology.

It's the gap.

The AI Version of This Problem

Here's where this gets relevant to your Monday morning.

Every organization right now has some version of an AI policy. Or an AI committee. Or an AI working group that meets monthly to evaluate AI readiness and report back to leadership with a recommendation sometime in Q3.

Meanwhile the person in the call center — or the operations team, or the finance department, or the project management office — has already figured out that ChatGPT cuts their weekly report from two hours to twelve minutes. They're doing it anyway. On their personal device. Outside the approved toolset. Without telling anyone.

Because the gap between finding the solution and getting permission to use it is so wide that waiting isn't rational.

So the AI strategy isn't happening in the committee room. It's happening in browser tabs that nobody is tracking. On personal phones. In the fifteen minutes before the 9am meeting starts.

Your people already have an AI strategy. You just don't know what it is.

What IT Security Gets Right and Wrong

IT security concerns about AI are real. I want to be clear about that.

Data privacy matters. Compliance matters. The risk of sensitive information entering a model that wasn't designed to hold it — that's a legitimate concern that deserves a legitimate answer.

What it doesn't deserve is being used as the permanent organizational excuse to avoid the conversation entirely.

"We can't use AI because IT security."

That's not a policy. That's a placeholder for a decision that nobody has made yet because making it requires understanding something unfamiliar — and that's uncomfortable for people who are used to being the ones with the answers.

The organizations getting this right are not the ones that said yes to everything with no guardrails. They're the ones that sat down with IT security and said: what specifically are we protecting, what tools meet that standard, and how do we get our people access to something real within ninety days.

That's a solvable problem. It just requires someone willing to solve it instead of defer it.

The Culture That Closes the Gap

Here's what MBNA understood that most organizations don't.

The people closest to the problem have the best information about how to fix it. Leadership's job is to close the distance between those people and the resources to act — not widen it with process and approval layers that create the illusion of control while producing the reality of paralysis.

Every manager and officer on the floor four hours a month wasn't a charity initiative. It was intelligence gathering at the source. When I found the broken field edit, I didn't need to write a report about it because I was already talking to Tony. The information gap was zero.

That's the standard.

Not the governance committee. Not the quarterly AI readiness report. Not the policy document that takes eight months to ratify and is already outdated when it's published.

Direct contact. Direct access. Problem found, solution delivered.

That was 1995. The tools are dramatically better now. The principle hasn't changed.

What This Looks Like Today

You don't need to recreate 1995. You need the same outcome with different tools.

Close the gap between the people who find problems and the people who can fix them. That means:

Leaders who are genuinely curious about what AI their people are already using — without making them feel like they're in trouble for finding solutions.

IT security teams who are asked to find a path forward, not just identify risks. Risk identification without solution architecture is just expensive worry.

Managers who model AI curiosity instead of AI avoidance. Your team is watching what you do, not what the policy says.

An approved toolset that includes something actually useful — not just Co-pilot with the guardrails locked to the point of uselessness — so people stop doing it on their personal devices and start doing it in a way you can see and learn from.

And some version of Tony. Someone who picks up the phone. Someone who tries it on the next call.

The problem finding is already happening.

The question is whether you're building an organization that can act on what it finds — or one that schedules it for Q3.

*Next: The 30-day AI experiment you can run right now. No budget. No permission. No transformation program required. Just a problem, a tool, and enough curiosity to see what happens.*

*2weekAI deploys AI that actually works — in 2 weeks, at a fixed price. Dave has delivered AI solutions across Fortune 100 enterprises and growing businesses alike. No transformation program required. [Book a discovery call.]*