Northline Strategy

08/19/2026

Forty minutes. Three executives, one governance committee, and a marketing AI pilot that needed legal review.

The policy was clear, and nobody disagreed with it. The problem was that each executive had a fundamentally different idea of what “acceptable risk” meant for the company.

One was thinking about reputational exposure. Another was thinking about regulatory liability. The third was thinking about operational disruption.

So they spent forty minutes debating the word “risk” when a shared definition could have ended the meeting in five.

This is something I see organizations get backward with governance. They focus on writing the policy, establishing the committee, and creating the review process. Those things matter, but they can’t manufacture shared understanding that doesn’t already exist.

Good governance starts earlier. Decision rights are understood before the hard decisions arrive. Escalation paths exist before conflicts do. Leaders have enough shared language around risk and purpose that a new situation doesn’t require them to renegotiate the fundamentals every time.

Governance works when it builds shared clarity.

08/19/2026

08/19/2026

Five leaders can leave the same meeting feeling aligned — and walk away with five different interpretations of what happens next.

That’s not an ex*****on problem.

It’s an alignment problem.

DM me ALIGNMENT and I’ll send you a simple exercise to help surface it.

08/17/2026

The pilot exceeded every metric. Adoption was high, the team loved it, and leadership was impressed. The data looked exactly like what you’d want to show a board.

The organization around it was never designed to absorb what it proved.

A pilot operates under unusually good conditions. It has concentrated leadership attention, a carefully selected team, and narrowly defined objectives. Success depends heavily on the quality of the solution itself.

Enterprise adoption introduces something different. Local workflows intersect with enterprise governance. Departmental priorities compete for shared resources. Questions about accountability, coordination, and ownership start surfacing at precisely the moment leadership believes everything is working.

That friction doesn’t necessarily mean the technology failed. It may mean the technology succeeded far enough to expose what the organization wasn’t designed to handle.

This is where AI stops being a traditional technology project and starts becoming an operating model question.

A successful pilot should answer two things: Did the technology create meaningful value? And what needs to change about the organization for that value to become part of everyday operations?

You can have dozens of successful pilots and still not have enterprise capability.

08/06/2026

How do we actually work with organizations? Not a rigid process but a philosophy, built around five stages: Observation, Alignment, Design, Implementation, Advisory.
Curious which stage your organization is in?

08/04/2026

In Austin next week.

From August 10–13, I’ll be in Austin for the CSPEN Annual Conference, where I’ll be presenting on how organizations can move beyond AI adoption and begin designing the capabilities needed to make it part of how they actually operate.

Conferences are always valuable, but some of the best conversations happen outside the sessions.

If you’ll be at CSPEN—or simply happen to be in Austin that week—I’d love to connect. Whether you’re leading AI initiatives, thinking through organizational transformation, or just interested in where leadership and technology are headed, I’m always glad to make time for a good conversation.

If you’re around, send me a message. Let’s find time for coffee.

08/04/2026

We just published the first in a new series of Executive Briefs: Executive Advisory.
The short version: most organizations don’t have a strategy problem. They have an organizational design problem, and AI is simply making that visible faster than ever before.
Read the full brief: https://northlinestrategy.co/insights/

08/03/2026

Most organizations that tell me they have AI governance actually have an AI policy.

Those are not the same thing.

A policy tells people what they're allowed to do. It sits in a shared drive. It gets reviewed once a year, maybe. It covers acceptable use, data handling, a few risk categories. It's usually well-written. And it is almost always disconnected from how decisions actually get made inside the organization.

Governance is different. Governance answers the question that comes after the policy gets published: when a new AI use case shows up on a Tuesday afternoon, who decides whether it moves forward, what criteria they use, and what happens if the answer is no?

I've reviewed frameworks where the policy was thorough and the decision path was completely informal. Leaders were making real calls about AI adoption based on hallway conversations and whoever happened to be in the room. The policy existed. The structure to act on it didn't.

That gap is where organizations get exposed. Not because the policy was bad, but because nobody built the governance layer underneath it. Decision rights, escalation paths, review rhythms, accountability after the decision gets made. That's the infrastructure most teams skip.

And honestly, it makes sense that they skip it. Writing a policy feels productive. Building governance requires leadership to sit with uncomfortable questions about who actually has authority and how decisions flow. That's harder. It's also the only part that changes behavior.

If your organization has an AI policy and still can't answer who owns the next AI decision, you have documentation. You don't have governance yet.

07/30/2026

Revisiting a piece from earlier this year that's aged into the current conversation.
Capability has been commoditized. Readiness hasn't.
Twelve months ago, deploying enterprise AI for document review or customer interaction required real technical investment. Today it's off-the-shelf, configurable by a reasonably technical product manager, and priced at a fraction of custom development.
That's not a problem. That's the point.
The competitive bottleneck moved. It's no longer access to capability — it's the speed at which an organization can redesign itself to use what's already available. Most leadership teams are still treating AI adoption as a procurement question: which tool, which vendor, which security review. The harder question is structural — does this organization's decision-making, governance, and incentive structure actually allow it to change how work happens?
That distance is the Acceleration Gap. It isn't a technology lag. It's an organizational one.

07/28/2026

A pattern worth naming before it becomes a bigger problem.
Ask five people on the same leadership team to describe the company's AI strategy in a sentence. Most will give you something reasonable. Ask a harder question — what has the organization explicitly decided not to do with AI, and why — and the answers diverge fast.
That's not disagreement. It's polite agreement followed by inconsistent interpretation. Everyone sounds aligned until the decisions start getting made, and "aligned" turns out to have meant five different things.
If every team can make the strategy sound reasonable in a different way, the organization doesn't have alignment. It has translation risk.