Inside the cockpit of a fighter jet, decisions happen under conditions no system can fully predict. A pilot is processing signals from everywhere: radar, instruments, communications, the movement of another aircraft. There isn’t time to gather perfect information or reconcile every data source. You observe what you can, orient yourself to what it means, decide what to do, and act.

‍

That cycle - Observe, Orient, Decide, Act - became known as the OODA loop, developed by Air Force strategist John Boyd while studying why American F-86 Sabre pilots achieved a 10-to-1 advantage over Soviet MiG-15s during the Korean War, despite the MiG having stronger technical specifications.

‍

Boyd’s insight wasn’t simply about moving through the loop faster. It was about maintaining control of the decision loop itself. Decisions had to remain grounded in evidence, owned by the people making them, and evaluated through the outcomes they produced. Each action generated new observations, allowing pilots and organizations to refine their understanding and improve the next decision.

‍

The advantage wasn’t speed alone. It was the ability to learn and adapt through the decision cycle. That idea stayed with me throughout my career and, today, is the foundation of Axonis Decision Intelligence.

A Career in Complex Decisions

I began working on command-and-control systems in the late 1980s and later moved into intelligence platforms used across U.S. and coalition operations. My role was rarely about building a single system. It was about connecting many systems, systems owned by different organizations, operating under different rules, and containing different pieces of information.

‍

In those environments, the challenge was never simply collecting data. There was always plenty of data. The real challenge was making decisions when the information was incomplete, fragmented, and constantly changing.

‍

Anyone who has worked in intelligence knows the world rarely presents clean answers. Most of the time, it’s gray. Leaders are making decisions with partial information, conflicting signals, and limited time. And they have to do it anyway.

‍

One lesson became clear early on: the information you need rarely lives in one place.

‍

Different organizations hold different parts of the picture. Some data can’t be shared freely. Some systems were never designed to integrate. In coalition environments, partners may each hold critical information but operate under strict rules about how it can be accessed.

‍

Centralizing everything into one system sounds appealing, but in practice, it rarely works. By the time you consolidate the data, the situation has already changed.

‍

So the real challenge becomes something else entirely: How do you make coherent decisions across distributed information without forcing everything into a single place?

‍

If that problem sounds familiar, it should, because it’s exactly what most enterprise businesses face today.

‍The Fog of Business

Military strategists often talk about the “fog of war” the uncertainty and friction that surround real-world operations. But the same phenomenon exists inside organizations.

‍

Call it the fog of business.

‍

Important decisions are rarely made with perfect information. The data needed to understand a situation is scattered across systems, teams, and policies. Finance may see one signal. Operations another. Security or compliance may hold critical context that no one else can access directly.

‍

Individually, each perspective makes sense. But when the information never comes together, the decision becomes fragmented.

‍

This is where many organizations struggle today. AI can generate insights and recommendations quickly, but when it comes time to act, something essential is often missing: the context behind how the decision was formed.

• What evidence was used?
• What signals mattered most?
• What policies shaped the outcome?
• Who ultimately approved the decision?

Without that context, organizations cannot evaluate the decision afterward, nor can they improve the next one.

‍

The decision loop breaks.

The Closed Loop Problem 

In complex environments, many OODA loops need to run simultaneously across teams and systems, each working toward the same objective but often with different information and constraints.

‍

And in most organizations, that process remains largely invisible. Decisions are made, actions are taken, but the signals, evidence, and reasoning behind them are rarely preserved in a way that can be revisited or learned from.

‍
LLMs collapse Observe + Orient + Decide into a black box.  They produce outputs without: preserved evidence, traceable reasoning and accountability. That creates a closed, un-auditable OODA loop.

‍
Many modern AI systems, particularly large language models, appear to accelerate decision-making. But in doing so, they often collapse the decision loop into a single opaque step. Observation, orientation, and decision are compressed into a model output. The reasoning is not preserved. The evidence is not explicitly tied to the outcome. The decision cannot be meaningfully audited or revisited.

‍

The loop closes and what looks like speed is often a loss of control.

‍

When the decision process becomes observable, something important changes. Organizations can see how their decision systems actually operate, where they slow down, where evidence is thin, and where outcomes improve.

‍

Over time, this creates a persistent record of the decision cycle, what we call an Evidence Ledger. It becomes a form of institutional memory, allowing organizations to revisit past decisions, refine how they interpret signals, and continuously improve how decisions are made.

‍

Our goal isn’t just to make one good decision. It’s also to improve the next one. Which means the decision loop must remain visible.  So, how do we bring the decision loop into AI-assisted systems?

From Boyd’s Loop to Decision Intelligence

When we began building Axonis Decision Intelligence, Boyd’s framework was already top of mind. The OODA loop isn’t just a philosophy of decision-making; it describes a structure for how decisions actually unfold.

‍

Observe. Orient. Decide. Act.

‍

Each phase produces information that feeds the next cycle. Once you start thinking about decisions this way, it becomes clear that the loop can be implemented architecturally.

‍

Axonis Decision Intelligence encodes that structure directly into the platform. Every phase - observation, orientation, decision, and action - creates traceable context that feeds the next cycle.

‍

Here is how the OODA loop translates into ADI.

‍

‍

What Boyd described as a strategic framework is the underlying operational system for Axonis Decision Intelligence.

• Signals surface what matters.
• Evidence blocks build orientation.
• Editions capture the decision.
• Attested actions close the loop and generate the next observation.

The loop doesn’t disappear inside a model. It becomes visible, accountable, and continuously improvable.  

Implicit Guidance and Control

One of Boyd’s most important ideas was what happens between the phases of the loop. He called it implicit guidance and control.

‍

When an organization shares the same mental model of a situation, people don’t need explicit instructions for every action. They understand the objective, the context, and the boundaries of the mission. That shared understanding allows teams to act independently while still moving in the same direction.

‍

In military operations, this allows people at the edge of an operation to make decisions without waiting for centralized commands. In organizations, it means teams can move forward without layers of approvals slowing everything down.

‍

The key is shared orientation.

‍

When people understand how decisions should be made and what evidence matters, they can act confidently while remaining aligned with the broader mission.

‍

That concept influenced how we designed Axonis Decision Intelligence.

‍

Within the platform, Packs help establish this shared orientation. Profile Packs define who is involved in a decision. Domain Packs define what the underlying data means. Accountability Packs clarify what constitutes a completed decision. Experience Packs shape how information surfaces during investigations.

‍

These Packs don’t dictate decisions or enforce rigid workflows. Instead, they establish the context that helps people understand what “good” looks like.

‍

Boyd understood that effective organizations don’t win by centralizing decision authority. They win by distributing understanding so that people across the organization can act within a shared framework.

Back to the Cockpit

Boyd wasn’t really describing dogfights. He was describing how organizations make decisions in complex environments.

‍

Observe the situation.
Orient to the evidence.
Decide what to do.
Act.

‍

Then repeat the cycle, learning from each outcome.

‍

The advantage doesn’t come from having the most data or the most sophisticated tools. It comes from maintaining control of the decision loop.  Boyd also recognized that real organizations do not operate through a single decision loop. Many OODA loops run concurrently across teams and systems, each working toward the same outcome.

‍

Artificial intelligence is accelerating analysis across every industry. But the real challenge for enterprises isn’t generating more answers.  

‍

It’s ensuring that decisions remain grounded in evidence, owned by accountable humans, and preserved in a way that allows organizations to learn and improve over time.  That’s what Axonis Decision Intelligence is about. It maintains control of the decision loop, enabling you to make the next decision better than the last.

‍

And in many ways, it brings us right back to the cockpit where the importance of that loop was first understood.

‍