Latest from moring
Technical guidance, research and product updates to help you build and govern AI agents as enterprise adoption accelerates.
Observability for AI Agents: Metrics, Traces, and Guardrails That Matter
Metrics say whether an agent works, traces explain why it decided, guardrails bound what it may do, and all three write to one shared record.
Ashvath NarayanHow should a bank use AI agents to investigate reconciliation breaks?
Scope a reconciliation agent to evidence: which causes it may name, what it must never do, what the controller receives, and what is kept for the next audit.
Arya SharanWho is accountable when an AI agent makes a bad decision?
Why "a human in the loop" is not an answer, and the four roles to record on every agent run: who initiated, which agent acted, who owns it, who approved.
Arya SharanLatest blogs
How should a bank use AI agents to investigate reconciliation breaks?
Scope a reconciliation agent to evidence: which causes it may name, what it must never do, what the controller receives, and what is kept for the next audit.
Arya SharanWho is accountable when an AI agent makes a bad decision?
Why "a human in the loop" is not an answer, and the four roles to record on every agent run: who initiated, which agent acted, who owns it, who approved.
Arya SharanWhat is AI agent orchestration, and how do you govern it from one place?
The three orchestration patterns enterprises deploy, why traditional governance tools fail at that layer, and five pillars for governing it from one place.
EllakkiaaEnterprise AI Architecture: Where an AI Control Plane Fits
Where an AI control plane sits in the enterprise stack, what it does on each agent action, how it relates to platforms you run, and whether to build or buy.
Ashvath NarayanAgentic AIOps: Who Governs the Ops Agents Now Acting on Their Own
Ops agents now restart services and roll back deploys on their own. The governance questions every deployment must answer, and where today's approaches break.
EllakkiaaObservability for AI Agents: Metrics, Traces, and Guardrails That Matter
Metrics say whether an agent works, traces explain why it decided, guardrails bound what it may do, and all three write to one shared record.
Ashvath NarayanFrom Pilot to Platform: A 90-Day Rollout Plan for Enterprise AI Agents
Why agent rollouts stall after the pilot, and a 90-day plan in three phases: standards as code, routing and secrets gating, then telemetry and a second team.
Balaji NagarajWhy MCP Exists: Connecting AI Agents to Tools and Enterprise Systems
Why a bigger context window doesn't solve live system access, how integration code outgrows the agent, and what MCP actually standardizes.
Vignesh NagarajanWhat an AI Control Plane Is, and the Three-Question Test for Evaluating One
What an AI control plane is in plain terms, and the Deny-Attribute-Replay test for evaluating one, plus where the tools you already own stop short.
Ashvath NarayanTaking MCP to Production: Tool Design, Identity, Gateways, and Governance
What changes when a local MCP server becomes shared infrastructure: workflow-shaped tools, context budget, on-demand loading, hosting, identity and governance.
Vignesh NagarajanAI-DLC vs AI-SDLC: What's the Difference?
AI-SDLC bolts AI onto the lifecycle; AI-DLC governs it. The seven-stage workflow spine, what a native audit trail contains, and how to measure maturity.
Balaji NagarajNo posts in this category yet.