The power of generative AI took the business world by surprise. It wasn’t until the release of ChatGPT that many executives truly appreciated the seismic impact of these large language models (LLMs), and many companies were left scrambling to keep up. As we enter what’s likely to be a period of permanent revolution, during which GenAI’s capabilities will progress much faster than businesses will be able to adapt, companies simply can’t afford to sit and wait. The next leap in AI—autonomous agents—could enter the mainstream in the next few years and promises to be even more transformative than today’s LLMs.

Although most current LLM-based applications change how information is gathered and delivered, they stop short of operating independently. Some can automate specific tasks, but they still require a human to input a series of prompts and monitor the output. In contrast, autonomous agents—which are in part made up of LLMs—will be capable of redesigning and automating entire workflows. They plan how to execute tasks end to end, iteratively querying LLMs (through application programming interface (API) calls, where one application requests data or services from another), monitoring output, and using other digital tools to accomplish a given goal. As we discuss in examples below, autonomous agents could be used to design, execute, and refine entire marketing campaigns or undertake R&D testing through at-scale simulation. Autonomous agents are, in effect, dynamic systems that can both sense and act on their environment. In other words, with stand-alone LLMs, you have access to a powerful brain; autonomous agents add arms and legs.

With stand-alone large language models, you have access to a powerful brain; autonomous agents add arms and legs.

The arrival of autonomous agents into the mainstream isn’t far off. Today’s agents still lack the controllability and predictability needed for widespread use, but technology firms are making constant improvements. OpenAI’s recently announced custom bots are a clear step in this direction; they are able to use external APIs to find specific information or to carry out simple actions like assisting with an e-commerce purchase. Companies should start preparing for wide-scale adoption of autonomous agents today by adjusting their generative AI strategic planning—including their technology architecture, workforce planning, operating model, and policies—to ensure their transformation roadmap is robust and ready.