The more autonomous AI becomes, the more human strategy matters

Over the past few years, discussions around AI have largely focused on its growing capabilities. From content generation and information summarization to data analysis, code writing, and completion of complex tasks, AI has advanced remarkably. With the emergence of AI agents, the handover of not just tasks but portions of workflows and decisions is becoming more common.

This progression has understandably sparked questions about the future role of humans as AI becomes more autonomous. However, the real change lies in the nature of the tasks people need to perform. When AI excels at execution, people must concentrate on the system itself, asking crucial questions. What goals are we aiming for? What data does the technology require to make a wise decision?

How should tasks flow from one to another? When can AI make a decision on its own, and when does human intervention become necessary? While companies already tackle these questions, AI raises the stakes due to the speed and scale at which decisions can be made. Consider a marketing team, for example. AI can now analyze audiences, create content, adapt it for various channels, verify it against requirements, and assess results.

It's not hard to envision these capabilities seamlessly connected so that one action automatically triggers the next. The key lies in coordination. Despite all functioning as designed, the result might still be a poor outcome. An audience model might identify likely responders without considering long-term customer development. A content agent could generate click-worthy content inconsistent with the brand's long-term vision.

A performance evaluator could recommend reasonable actions based on a metric that doesn't reflect the overall business objective. These aren't necessarily AI failures but rather failures in design. The solution lies in people understanding the broader business context and ensuring that individual decisions across a workflow support the same goal.

Large enterprises have seen this before, with the integration of various specialized technologies for specific needs. While each solved real problems, coordinating these systems became a challenge. Employees effectively became the integration layer, bridging gaps between disparate platforms. As companies adopt AI across functions, they may inadvertently recreate this problem, only more rapidly with technology capable of independent decision-making.

Companies can easily adopt AI tools or build agents for specific problems, but as they proliferate, coordinating them becomes critical. The individual components may be more intelligent than previous software systems, but coordinating them is even more consequential, making strategy more important than ever. Finding the right balance is essential.

Human strategy plays a crucial role in designing this larger system, but it doesn't mean putting a person behind every AI decision. Requiring approval for every output would undermine the purpose of autonomous systems. Companies must determine where AI can operate independently, where monitoring suffices, and where direct human involvement is necessary.

These boundaries will differ across organizations and even within the same company. A routine decision with clear rules and limited consequences may need minimal intervention, while decisions affecting customer relationships, brands, significant spending, or strategic direction require more scrutiny. At times, the system may need to recognize unusual events and involve a person.

Determining these boundaries and designing workflows around them may prove more critical than the specific model or agent chosen by the company. Technology will continue to evolve rapidly. New, superior models may emerge quickly, and capabilities that seem unique today may become common. Companies cannot rely on a specific model for a lasting advantage, especially as gaps between competing technologies shrink.

What will be harder to replicate is the strategic thinking that surrounds the technology – an organization's understanding of its customers and operations, the design of its workflows, and the context provided to its AI systems.