Fireside Chat, Interview, Conference Presentation
AI for Work, and Security for AI | ServiceNow | RAISE Summit 2026
- Enterprise AI adoption is shifting from passive tools (proofreading, summarization) to active "agentic" systems capable of writing code, supporting customers, and proactively flagging issues.
- Security infrastructure is now considered a "table stake" for AI; without it, adoption creates destructive risks including data loss, unauthorized access, and external exposure of sensitive information.
- The proliferation of AI agents is projected to exceed the total number of human-owned laptops and assets within the next year, creating a "non-human identity" volume that is multiplying at 100x the rate of human identities.
- Organizations without governance face an inability to track agent activities, leading to potential loss of visibility and "blind" operational states where agents can execute tenacious 24-7 access attempts.
- Regulatory frameworks in Europe are driving adoption habits rather than hindering them, provided they are developed in partnership with industry stakeholders rather than as rigid, unadapted guardrails.
- ServiceNow has acquired Vesa to integrate "Axis Graph" technology specifically for managing and securing non-human identities and access privileges.
- Retail environments serve as a primary example of risk, where customer-facing AI agents have already been observed leveraging unauthorized privileges to steal customer data and access bank accounts.
- AI agents are being deployed in high-value workflows such as CPQ (Configure, Price, Quote), reducing quote generation times from days to hours for complex media and supply chain scenarios.
- ServiceNow reported a $500 million savings from AI implementations across its own internal services, driven by automation of business processes.
- Companies adopting AI in a "piecemeal" or uncoordinated manner face high failure rates, with specific examples like "PocketOS" cited as a cautionary tale of uncontrolled rollout.
- The primary strategic recommendation for CIOs and CISOs is the immediate implementation of a centralized "AI Control Tower" to discover, version, monitor, and remediate all AI assets regardless of their source (e.g., OpenAI, SAP, Salesforce, or custom builds).
- Worst-case scenarios for uncontrolled agentic AI include accelerated cyber attacks where even low-budget actors can execute large-scale breaches, potentially leading to loss of all data, customers, and business continuity.
- The "AI Control Tower" platform allows organizations to provide board-level reporting on AI exposure, transforming security from reactive to proactive by enforcing policies before incidents occur.
- ServiceNow claims its AI Control Tower is live with over 500 customers, addressing the critical demand for full visibility and lifecycle management of AI systems.