Proofpoint has introduced a SOC Analyst Agent powered by OpenAI Daybreak models, aimed at automating the investigative work between a security alert and a human response decision.
Analysts can use natural-language prompts to investigate alerts, logs, data-loss-prevention events and user-risk signals. Proofpoint says the agent returns structured findings, supporting evidence and recommended next steps, and can also schedule recurring investigations.
The autonomy boundary is deliberately narrow. The current agent can investigate and recommend, but it does not autonomously contain threats, change accounts or execute other consequential remediation.
The product is in private preview with beta customers, with general availability targeted for the end of the third quarter of 2026.