Agentic AI Matures to "Digital Worker"
The AI paradigm is shifting from analytical tools to agentic AI, which functions as an autonomous "digital worker." These agents can now perform complex, multi-step tasks by integrating with hundreds of APIs, executing code, and querying databases, significantly compressing innovation and operational cycles.
- Defense technology company Anduril's Lattice for Mission Autonomy software enables a single human operator to control teams of unmanned systems across air, land, and sea. In the US Army's EDGE23 exercise, the software allowed an integrated team of uncrewed aircraft to locate, identify, and destroy a surface-to-air missile site. - The Department of Defense's 2023 "Data, Analytics, and Artificial Intelligence Adoption Strategy" outlines the goal of accelerating AI adoption to achieve superior battlespace awareness and faster, more resilient kill chains. This formal strategy builds on pathfinder programs like Project Maven, an initiative started in 2017 to fast-track the integration of machine learning for analyzing surveillance data in combat theaters. - The technical stack for agentic AI typically uses a Large Language Model (LLM) like GPT-4 as a reasoning engine, combined with orchestration frameworks such as LangChain or AutoGen to plan and execute multi-step tasks. These frameworks allow the "digital worker" to use external tools and APIs to interact with different software and data sources. - Venture capital funding for robotics startups saw a significant surge, with companies raising over $4.2 billion in early 2024. Major deals included AI robotics company Figure raising $675 million in a Series B round with backing from Microsoft, Nvidia, and Amazon, and Physical Intelligence, a startup building AI models for robots, securing a $400 million Series A. - In commercial robotics, humanoid robots are moving into active deployment; Agility Robotics has deployed a small fleet of its Digit humanoids at a Spanx logistics facility in Georgia. Similarly, Figure AI is shipping its Figure 02 system to its first paying customers after testing its use in automotive production with BMW Group. - Enterprise software is shifting from traditional Robotic Process Automation (RPA) to more autonomous agentic systems from companies like UiPath, ServiceNow, and Microsoft. These platforms create "digital workers" capable of handling complex, end-to-end business processes, such as financial monitoring or supply chain optimization, with less need for human intervention.