Agentic Explainability at Scale: Between Corporate Fears and XAI Needs

TL;DR

This paper discusses the challenges of scaling agentic AI in enterprises, highlighting governance concerns, and proposes explainability techniques and a prototype tool to improve transparency and trust.

Contribution

It introduces design-time and runtime explainability methods and a prototype Agentic AI Card to address governance fears in large-scale agentic AI deployment.

Findings

Identifies governance concerns related to agentic AI at scale

Proposes explainability techniques for agent configuration and decision processes

Develops a prototype Agentic AI Card for deployment reassurance

Abstract

As companies enter the race for agentic AI adoption, fears surface around agentic autonomy and its subsequent risks. These fears compound as companies scale their agentic AI adoption with low-code applications, without a comparable scaling in their governance processes and expertise resulting in a phenomenon known as "Agent Sprawl". While shadow AI tools can help with agentic discovery and identification, few observability tools offer insights into the agents' configuration and settings or the decision-making process during agent-to-agent communication and orchestration. This paper explores AI governance professionals' concerns in enterprise settings, while offering design-time and runtime explainability techniques as suggested by AI governance experts for addressing those fears. Finally, we provide a preliminary prototype of an Agentic AI Card that can help companies feel at ease deploying agents at scale.