Back to Insights
Future of Work6 min read

Why Humans + AI Recursive Self-Improving Decisions Will Reshape Business

The next phase of business will be centered on Humans + AI recursive self-improvement.

At the heart of every organization lies a single activity: making decisions. Operational decisions about inventory and scheduling. Strategic decisions about markets and investments. The countless micro-decisions that accumulate daily into what we call organizational performance.

Improving decision quality is where the greatest value lies. Speed and cost matter too. But the organizations that make consistently better decisions will decidedly outperform.

We are crossing a threshold where we can build systems that learn from decisions at every level of the organization, capturing not only what was decided but how it was decided, in what context, and what followed.

From systems of record to systems of intelligence

One of the ideas that has most impacted major software company valuations over recent months is the shift from systems-of-record to systems-of-agents built on context graphs.

Traditional enterprise software such as ERP, CRM, and marketing suites, are fundamentally systems of record. They capture data that feeds into decisions. They are repositories, not reasoning engines. A context graph is something profoundly different: a living map of how organizations function and how decisions are made.

Systems of record tell you what happened. Context graphs can tell you why it happened, what the decision-maker knew at the time, what alternatives were considered, and what the downstream effects were.

The value of this vastly transcends any traditional software platform.

The opportunity is not delegation

The opportunity is not delegating decisions to AI.

It is building systems that draw on how decisions throughout the organization have been made, the context in which they were made, the outcomes from those decisions, and the lessons learned. This is fundamentally a Humans + AI proposition. Every significant decision in an organization is shaped by human judgment, context, relationships, and values. AI does not replace these. It augments them.

The configuration of humans and AI in decision-making will rapidly evolve. In some operational domains, AI will handle more of the routine assessment, freeing human attention for the judgments that require experience, ethics, and situational awareness. In strategic domains, AI will surface patterns and possibilities that no individual or team could identify unaided.

Most importantly, we can now build systems that capture the implicit as well as explicit decision process. The reasoning that was never documented. The contextual factors that shaped a judgment but were never entered into any system of record. The tacit knowledge that walks out the door when experienced leaders leave.

Recursive self-improvement as an organizational capability

In AI research, recursive self-improvement refers to systems that can enhance their own capabilities. Applied to organizations, the principle is even more powerful: a Humans + AI system that continuously learns from its own decision-making, refining not only the AI models but the human processes, frameworks, and judgment that drive outcomes.

Each decision becomes a data point. Each outcome becomes a lesson. Each lesson feeds back into the system. This is a flywheel, and organizations that build it first will compound their advantages rapidly.

As I have explored in the context of the explore-exploit dilemma, intelligence itself is rooted in the capacity to learn from choices and adapt. Organizations can now externalize and systematize this capacity in ways that were previously impossible.

There is of course the potential for misuse. Systems that capture decision processes raise questions about surveillance, autonomy, and organizational trust. These are legitimate concerns requiring thoughtful governance. But applied well, there can be immense value in organizations that genuinely improve their collective intelligence over time.

The companies building the infrastructure for decision intelligence, the context graphs, the agent systems, the feedback loops, are laying the foundation for what business will become.

The organizations that recognize and act on this earliest will build advantages that compound with every decision they make.

Ross Dawson

About Ross Dawson

Ross works with leadership teams to design organizations, business models, and decision-making structures for a world shaped by AI. His work focuses on how Humans + AI reshape strategy, work, and leadership—helping organizations move beyond experimentation toward real, scalable impact.

View Full Profile →