Summary

The rapid adoption of enterprise AI, particularly autonomous agents, is creating a much larger and more complex attack surface for organisations. Unlike chatbots, agents can access systems, carry credentials and take independent action, making them vulnerable to manipulation through techniques such as prompt injection and jailbreaks. The immediate challenge for companies is visibility, as many do not know how many agents are operating, what they can access or whether they are behaving improperly. Cybersecurity vendors are therefore developing tools to identify, monitor and control agents, with identity, endpoint, network and exposure management expected to be key areas of investment. At the same time, increasingly capable AI models are allowing attackers to discover and exploit software vulnerabilities faster and at greater scale. The traditional gap between vulnerability disclosure and exploitation is collapsing, forcing organisations to modernise legacy systems, adopt virtual patching and remove weaknesses between security products. Marker expects cybersecurity growth to begin reaccelerating from the fourth quarter, with the larger impact emerging in 2027 as enterprise sales pipelines convert into revenue. While cybersecurity shares may remain volatile in the near term, the firm sees this as a generational, multi year opportunity for the sector.

Topics