Most organisations do not have an AI strategy. They have a collection of pilots.
We determine where AI creates value, make the hard calls on what belongs to AI and what does not, and embed governance into the system from the start. The board gets a defensible position. The CFO gets an answer to where the spend is going.
WHY ORGANISATIONS COME TO THIS
Three situations we see again and again.
-
01
The prototype graveyard
Pilots launched, budgets spent, nothing in production. The organisation has learned that AI is hard to ship, but not yet why. The answer is almost always that the strategy came after the prototype, and the governance never came at all.
-
02
The board wants a position
Leadership knows the organisation needs to have a view on AI. The pressure is real, from the market, from competitors, sometimes from regulators. What is missing is a structured answer that holds up in the room.
-
03
The CFO's question
Six months of AI spend. What has it produced? The question is fair and the answer is often uncomfortable. The engagement starts by making that question answerable, and then making sure it stays that way.
THE CORE WORK
AI strategy is not a technology decision. It is a value decision.
Before any architecture is drawn, the work answers three questions.
FIG · 01 · THREE QUESTIONS
-
01
Where does AI create value?
Not every problem wants a model. The strategy maps where AI genuinely adds value, and where deterministic code is cheaper, faster, and more testable. That call is made explicitly, for each part of the system, before anything is built.
-
02
Where does a human stay in the loop?
Human-in-the-loop is a governance decision, not an afterthought. Where it is required depends on the risk profile of the decision, the regulatory context, and what the EU AI Act requires for the system category. It is designed in, documented, and owned.
-
03
Who owns what?
Every decision point has an owner. Every AI output that feeds a human decision has a defined accountability chain. That is not bureaucracy. It is what makes the system defensible to an auditor, to the board, and under the EU AI Act.
GOVERNANCE EMBEDDED, NOT ADDED LATER
The most common governance failure is timing.
The system gets built, then someone asks about governance, and the answer is a document that sits next to the system rather than inside it.
At Imageplus, governance is part of the architecture from the start. Ownership, decision points, and human oversight gates are documented in BPMN alongside the process flows. The result is a system where the governance layer is not a parallel track. It is the same track.
BPMN 2.0 is the shared language across strategy, governance, and build. What is documented in the strategy phase becomes the architecture. What is built matches what was governed. The chain is unbroken.
HOW FAR IT GOES
A complete engagement. Or the start of more.
The strategy and governance work is a complete engagement in its own right. It produces a documented, validated AI position the organisation can act on, present to the board, and defend under audit.
Where the scope includes a build, it flows into a Forward Deployed Engineering mission or an AI lead fractional engagement. The strategy defines what gets built. The build matches what was governed.
Tell us where your AI programme stands today.
We will tell you whether this engagement is the right move, and what it would produce.
COMMON QUESTIONS
Asked before starting.
-
What if our AI project turns out not to be worth pursuing?
We will tell you. The point of AI strategy is to determine where AI creates value and where it does not. If the answer is that a project should not go ahead, that is a result, not a failure. It is also the most useful thing we can tell a CFO before the budget is spent.
-
What is the difference between AI and deterministic code?
Some problems want a model. Many are better, cheaper, and more testable as plain deterministic code. The strategy work makes that call explicitly, for each part of the system, before anything is built.
-
What is human-in-the-loop and when is it required?
Human-in-the-loop means a person stays in the decision chain at defined points, by design. Where it is required depends on the risk profile of the decision, the regulatory context, and what the EU AI Act requires for the system category. It is a governance decision, not an afterthought.
-
How does governance get embedded into the system?
Governance is designed into the architecture from the start, not bolted on afterwards. Ownership, decision points, and human oversight gates are documented in BPMN alongside the process flows. The result is a system that is defensible to an auditor and to the board.
-
How does this relate to the EU AI Act?
The EU AI Act imposes obligations on how AI systems are designed, governed, and overseen. Some requirements are already in force. The bulk applies from August 2026, with the remainder by 2027. Getting the architecture and ownership right from the start is the only way to meet those obligations without retrofitting governance onto a system that was never designed for it.
-
Can this engagement lead into a build?
Yes. The strategy and governance work defines what should be built and how. Where the scope includes a build, it flows naturally into a Forward Deployed Engineering mission or an AI lead fractional engagement.