The state of AI agents, in numbers (www.getreadyforagents.com)

Recent research from the Get Ready For Agents initiative reveals a striking uptick in AI agent-focused hiring, with nearly two-thirds of 3,164 tracked AI job postings aimed at building or operating agents, highlighting an urgent demand for these technologies. The data indicates a significant shift towards agent-centric roles, particularly within AI-native firms, but it also exposes a troubling lag in governance and oversight mechanisms for these autonomous systems. As organizations rush to deploy AI agents, only 23% are actively scaling them, pointing to a substantial gap between experimentation and practical application. The implications are profound for the AI/ML community, especially as 89% of teams developing agents have implemented observability, but only 52% are performing evaluations—an alarming 37-point gap. Furthermore, projections suggest that over 40% of current AI agent projects may be scrapped by 2027 due to rising costs and unclear business value. Meanwhile, notable successes, such as Klarna's AI assistant and JPMorgan's COiN platform, underscore the potential efficiency gains of these agents, though it remains evident that robust governance frameworks are essential for sustainable progress in this rapidly evolving field.