Asurion SOC Uses About 45 AI Agents to Monitor Security Data
· 2:37 · Watch on YouTube ↗
Summary
An Asurion SOC speaker describes deploying about 45 AI agents to review security data, run searches and flag data-source problems, while emphasizing context and verification to address AI hallucinations.
Key points
- The Asurion SOC has deployed about 45 AI agents, with the number possibly nearing 50.
- The agents continuously review data sources, run searches when needed and report problems with data sources.
- The speaker says AI hallucinations were a major challenge and stresses verifying responses rather than relying on them unchecked.
- Asurion worked with the Splunk AI assistant team on providing context and improving prompts to make responses more reliable.
- The speaker compares agents to interns: useful, but their work still needs to be verified.
Article Details
- Topics
- [00:16] Asurion SOC's goal of building a trusted AI platform
- [00:50] AI hallucinations and the need to verify responses
- [01:31] Using Splunk AI assistant and improving context and prompts
- [01:56] Deployment of about 45 AI agents working around the clock
- [02:05] Agents reviewing data, running searches and reporting data-source problems
- [02:21] Treating agents like interns in a trust-but-verify workflow