Success Stories: How Four Enterprises Safely Implemented AI Agent Identities

Visa’s Innovative Approach to Agentic Security

Visa’s president of technology, Rajat Taneja, recently presented at the VB Transform 2026 event, showcasing how Anthropic’s Mythos was applied to Visa’s payment network. By identifying and addressing minor weaknesses in the system, Visa was able to enhance its security measures. The company even made the harness used in the process open-source, allowing others to benefit from their findings.

While Visa’s engineering depth allowed them to take proactive steps in improving their security, many enterprises struggle to reach that level of preparedness. Research indicates that over half (53%) of enterprises have experienced a security incident or near-miss. Despite 65% of enterprises enforcing agent permissions at runtime, only 18% isolate their highest-risk agents, and a mere 8% combine enforcement with isolation.

Relying solely on provider-native controls for agentic security can widen the gap further. A recent study by VentureBeat Pulse Research revealed that 92% of enterprises default to using hyperscalers and AI platform providers for their primary security layer.

Discrepancy Between Satisfaction and Incident Data

Interestingly, research shows that enterprises tend to rate familiar tools higher, even if those tools have previously failed them. Data from a recent survey indicated that enterprises that experienced security incidents rated their satisfaction with security tools at an average of 4.39 out of 5. On the other hand, enterprises that had not encountered incidents rated their tools at 4.13. This discrepancy suggests that enterprises reward tools that prevent breaches with higher satisfaction scores.

One notable trend is the increasing importance of identity and isolation in security measures. While 49% of surveyed enterprises assigned each agent a managed identity, only 11 of those enterprises also implemented isolation measures. This highlights the need for a comprehensive, layered security strategy that includes both identity management and isolation.

Furthermore, enterprises that enforce scoped permissions at runtime without implementing isolation measures have a higher incident rate of 58%. This underscores the importance of combining enforcement with containment to effectively mitigate security risks.

Provider Lock-In and Future Plans

Provider-native platforms have become increasingly popular among enterprises, with a significant majority naming hyperscalers and AI platform providers as their primary security layer. This trend indicates a growing reliance on external providers for security solutions.

Despite high satisfaction levels with existing security tools, 74% of enterprises plan to replace these tools within the next 12 months. This suggests that enterprises are seeking more advanced and effective security solutions to address evolving threats.

Challenges and Opportunities in Agentic Security

As enterprises continue to deploy AI agents, it is crucial to prioritize the development of comprehensive security measures to manage these agents effectively. The data shows a clear need for a shift towards implementing isolation measures in addition to identity management to improve overall security posture.

By understanding the importance of combining enforcement with containment, enterprises can better protect their systems from security threats. The next wave of research will likely provide further insights into how enterprises are adapting their security strategies to address the containment gap effectively.

Conclusion

Visa’s proactive approach to enhancing agentic security highlights the importance of continuous innovation and improvement in cybersecurity measures. By addressing vulnerabilities and leveraging advanced technologies, enterprises can strengthen their security posture and better protect their systems from emerging threats.

It is clear that a comprehensive security strategy that includes identity management, isolation measures, and enforcement controls is essential for mitigating security risks effectively. As the cybersecurity landscape continues to evolve, enterprises must prioritize investing in advanced security solutions to safeguard their valuable assets.