Enterprises struggle with multi-agent conflict as adoption surges, study finds

A clear majority of enterprise technology executives believe they have agent deployment under control, yet few know how to handle autonomous systems when they clash, according to a joint study by OutSystems, CIO Dive, and KPMG ^[1].

The research, "Engineering Business Outcomes," surveyed more than 850 IT decision-makers ^[1]. The findings show a distinct operational divide: 82% of surveyed leaders stated they can keep pace with governing and integrating AI agents ^[1]. However, only 26% reported being well prepared to resolve conflicts that emerge between multiple agents operating within the same environment ^[1].

OutSystems, a Lisbon-based company that provides a low-code software development platform for enterprise software and agentic workflows, emphasized the need for governance embedded at the structural layer ^[1]. Tiago Azevedo, CIO of OutSystems, said companies should build control into the platform rather than reviewing it later ^[1]. Michael Harper, a managing director at KPMG, also contributed findings to the research project ^[1].

The survey highlights an emerging operational bottleneck for companies moving beyond isolated generative copilots toward coordinated agent teams, where automated decisions interact and risk creating logic deadlocks without deliberate orchestration rules ^[1].