Artificial intelligence in cybersecurity could improve efficiency in specific workflows like code scanning but it does not have the visibility, control or reliability to replace end-to-end security platforms, BofA Securities said in a Tuesday note.

Noting Anthropic's launch of Claude Code Security last week, BofA said the offering enables detection beyond simple pattern recognition, potentially improving accuracy on complex vulnerabilities.

While this is "meaningful," it is confined to the domain of preproduction code analysis, the note said.

"While AI-based tools offer an enhancement, they remain sensitive to prompt phrasing and lose context outside well-bounded tasks," BofA said.

Automated tools, powered by AI-reasoning models can enhance scanning-oriented features, BofA said, adding that cybersecurity companies have entered this segment via acquisitions, including CrowdStrike's (CRWD) purchase of Bionic and Palo Alto Networks' (PANW) acquisition of Bridgecrew.

Price: 353.31, Change: +2.98, Percent Change: +0.85