Notifications
Clear all
Topic starter
20/08/2026 5:52 pm
TL;DR: As organizations scale agentic AI into complex workflows, the security problem shifts from model quality alone to runtime governance, observability, and misuse resistance, according to Fiddler's analysis. The practical challenge is that autonomy expands prompt injection, hallucination, and data access risk faster than conventional ML metrics or static controls can track.
NHIMG editorial — based on content published by Fiddler: Developing Agentic AI Workflows with Safety and Accuracy
Questions worth separating out
Q: How should security teams govern agent tool calls in production?
A: Treat each tool invocation as a separate identity decision, not as a one-time permission granted to the agent.
Q: Why do agentic AI systems create more security risk than standard chatbots?
A: Agentic systems can turn model output into action, which means a bad instruction can affect code flow, tool use, and downstream state.
Q: What do security teams get wrong about prompt guardrails?
A: Teams often treat prompt guardrails as if they were authorisation controls, but they are only one layer of defence.
Practitioner guidance
- Define task-scoped agent identities Assign each agent workflow the minimum permissions required for a single business function, with explicit boundaries for read, write, and tool invocation.
- Enforce runtime policy at the tool layer Block or approve every external action through policy checks before the agent can call APIs, write records, or retrieve sensitive data.
- Build adversarial tests into release gates Simulate prompt injection, jailbreaks, retrieval poisoning, and unsafe tool requests before deployment.
What's in the full article
Fiddler's full blog covers the operational detail this post intentionally leaves for the source:
- Concrete examples of agentic workflows in media, finance, and research operations
- Practical monitoring ideas for hallucination, toxicity, and external attack detection
- A composite metric approach that blends LLM signals with business outcome measures
- Implementation considerations for LLMOps, human review, and scale planning
👉 Read Fiddler's analysis of developing agentic AI workflows with safety and accuracy →
Agentic AI workflows: are your safety controls keeping up?
Explore further
20/08/2026 6:31 pm
Agentic AI governance is becoming an identity problem, not just a model-safety problem. Once an agent can act across tools and systems, its permissions, approvals, and revocation lifecycle matter as much as the model itself. That makes the closest governance analogue NHI control, because the system needs scoped entitlement, runtime checks, and lifecycle accountability. Practitioners should treat agent behaviour as an access issue that belongs in IAM and PAM design.
A question worth separating out:
Q: How do organisations know if agent security controls are actually working?
A: Look for evidence that the platform can inspect traces, classify risky actions, and stop unsafe tool use before completion. Effective controls leave an audit trail that shows why the action was allowed or denied, and they reduce false positives enough that teams can trust them in production.
👉 Read our full editorial: Agentic AI workflows need stronger safety, accuracy and governance
