Keerthi Agnihotram
Articles by this author
Why enterprise AI is only as good as its search layer
Enterprise AI can only deliver reliable answers and actions when it can retrieve the right information and context. Based on a conversation with Forrester’s Rowan Curran, this blog explains why enterprise search and RAG are becoming foundational to agentic AI, covering data connectivity, retrieval quality, interoperability, evaluation, agent sprawl, and the shared search infrastructure enterprises need to move from isolated AI use cases to scalable, measurable outcomes.
The lesson enterprise AI needs to learn from the OpenClaw incident
The OpenClaw incident shows why instructions inside a prompt are not enough to govern AI agents. When a safety instruction disappeared during context compaction, the agent took actions it was explicitly told to confirm first. This blog examines what that failure means for enterprise AI and why access controls, confirmation gates, observability, audit trails, human oversight, and system-level guardrails need to be built into the platform itself.
Rethinking readiness: What AI is teaching enterprises about change
Enterprise AI readiness is becoming less about predicting what comes next and more about building an organization that can absorb continuous change. Drawing on conversations from Reimagine NYC, this blog outlines five foundations for adaptability: modular architecture, bounded autonomy, embedded governance, production-ready infrastructure, and absorptive capacity. Together, they help enterprises adopt new models, agents, and workflows without repeatedly rebuilding the systems around them.
Agentic AI evolution
Agentic AI is the latest stage in AI’s evolution from rule-based systems and chatbots to generative AI that can understand, decide, and act. This blog explains what defines an AI agent, the difference between agentish and agentic systems, types of agents, and six stages of enterprise adoption. It also covers practical use cases, governance requirements, implementation challenges, and the economics of deploying agents.