Enterprises tackle AI governance and security amid AI agent surge

IT and security leaders are confronting a new governance challenge as AI meeting assistants and autonomous AI agents become standard features in collaboration platforms. Companies report that AI‑driven transcripts create permanent records of sensitive conversations, raising questions about data ownership, retention policies, and access controls.

The rapid adoption of AI agents is also expanding the attack surface. Front‑line developer workstations are now gateways to cloud resources, prompting vendors such as Upwind Security to launch endpoint‑level AI sensors that correlate device activity with cloud actions. Experts warn that the window from vulnerability discovery to exploitation has shrunk to minutes, making real‑time visibility and identity management essential.

Across Europe, businesses are seeing the cost side of the trend. Gartner forecasts that by 2028 AI‑driven development costs could exceed average developer salaries, while a German study notes many firms are surprised by token‑based pricing. In the UK, regulators stress GDPR compliance and highlight the rise of “shadow AI” where employees use unapproved tools, prompting new guidance from the ICO and NCSC.

Vendors are responding with new governance frameworks that focus on data‑layer controls, auditability, and outcome‑based pricing. Salesforce introduced an AI service assistant and an outcome‑based Help Agent priced per resolved case, while a market‑wide ranking lists providers that can discover and classify sensitive data used by AI agents. The consensus is that robust policies, centralized AI inventories, and continuous monitoring are required to keep AI benefits from being eclipsed by security and compliance risks.