Tag: AI Governance
-
Agentic AI KPI: Straight-Through Rate
Part 2 of my KPI deep dives. An agent that never asks for help will post a great no-touch rate. Here is the partner metric that shows whether that autonomy was earned.
-
Agentic AI Cycle Time: Measure the Outcome, Not the Agent
Most dashboards time the agent. The customer feels the whole clock. Here is where to start and stop it, and three cases that show the difference.
-
Agentic AI: 5 Key Performance Indicators (KPIs) Besides Cost Savings
A large bank led its first agentic AI launch with time, not savings. Here are five KPIs that move a program from pilot to production faster than cost savings alone.
-
Agentic AI 101: From Punched Cards to intent.md
Agentic AI 101 starts with a punched card. In the 1970s only computer scientists could tell a computer what to do. Then programmers could, then business users could through screens IT built. Now the business user writes intent.md and an agent does the rest. Fifty years, one boundary, and a governance lesson at the end.
-
Top 10 Ways Agentic AI Delivers Value, Besides Cost Savings
Gartner now predicts 30 percent of employees replaced by AI will be rehired by 2029. I count ten ways agentic AI pays back that have nothing to do with headcount.
-
Agentic AI in Enterprise IT: 25 Things Retired by 2031
25 things AI agents retire in enterprise IT, from the service desk to seat-based SaaS licensing, with my dates and the controls to put in first.
-
What is Agentic AI Judgment Tier? Signature, Not the Steps
Agents now run more than 90 percent of their work steps without a person stepping in. The agentic AI judgment tier is what stays with a human: intent, done, gates, risk, controls and escalation. It sits with whoever owns the outcome.
-
Enterprise Applications to Lose $234 billion of spending to agents by 2030
Gartner puts $234 billion of enterprise application spend at risk from agents by 2030. I mapped 25 unnamed vendors into four phases, and the ones that survive own the record, not the screen.
-
Agentic AI Harness: Nobody Drives an Engine
An agentic AI harness is everything in the agent except the model. The engine is the model. The fuel tank, transmission, gears, brakes, key, governor and dashboard are the harness. This post maps all seventeen, one car part at a time.
-
AI Adoption in IT Services: Sell Judgment, Not Hours
Anthropic’s agents now log six hours for every human hour. For IT firms that bill by the hour and grow by hiring graduates, that changes pricing, staffing and who signs the work.
-
Agentic AI Pilot to Production: A Gate, Not a Graph
Two out of three agentic AI projects never leave the pilot stage. The firms that ship treat production as a gate you earn, not a graph you buy.
-
Agentic AI Without LLMs? It Flew a Spacecraft in 1999
Agentic AI without LLMs has run spacecraft, data centers and mortgage underwriting for decades. It works only inside a closed world a human built. Here is what the LLM adds, what it does not, and how a regulated enterprise wires the two together.
-
Agentic AI Harness Drift: Version the Harness Like a Model
You cannot claim change control because you pinned a model version. You can claim it when the harness itself is an inventoried, versioned, eval-gated artifact.
-
Agentic AI Contract Renegotiation: Renegotiate Before Renewal
Agents changed the work under your vendor contracts, but the invoice didn’t change. Agentic AI contract renegotiation resets price, risk and oversight before renewal.
-
Agentic AI Liability: Who Owns the Agent’s Decision?
An agent can follow every rule and still get it wrong. Here is who carries the loss, how far a vendor contract really helps, and the evidence that protects you.
-
Agentic AI Statement of Work: Outcome, Not Effort
What is an agentic AI statement of work? An agentic AI statement of work is the contract that says what an agent-run system must deliver and how acceptance is measured. It also sets the autonomy level and names who signs for each control. A traditional SOW buys effort, but this one buys an outcome with…
-
Agentic AI RACI Matrix: Governance Owns the Autonomy
What is an agentic AI RACI matrix? An agentic AI RACI matrix is a one-page chart for an agentic AI project. It names who is Responsible, Accountable, Consulted and Informed for every activity in the lifecycle. The chart runs from the first intent statement to the decision to expand or roll back autonomy. It differs…
-
Agentic AI Traces: Defend What You Can Replay
What are agentic AI traces? Agentic AI traces are the linked record of one agent run. They tie the first prompt, every model call, every tool call, and every permission decision to one ID, so the run can be replayed end to end. A log tells you something happened. A trace tells you what led…
-
Agentic AI Escalation Design: Name the Human First
What is agentic AI escalation design? Agentic AI escalation design is the decision, made before go-live, about who an agent calls when it is stuck, how it reaches them, and what happens if they do not answer. The harness enforces it, not the prompt. So an agent that pauses for a human it cannot name…
-
Agentic AI Audit Trail: Log the Handoff, Not the Answer
What is an agentic AI audit trail? An agentic AI audit trail is the harness’s record of every handoff, every tool call and every gate decision. The harness writes it, not the agent. It answers one question after the fact: who did what, on whose brief, and what did they see. So when the orchestrator…
-
Agentic AI SDLC Best Practices: Close the Loop Before You Leave
What are agentic AI SDLC best practices? Agentic AI SDLC best practices are the working rules for a lifecycle where an agent reads the code, plans, edits, runs commands and checks its own work. A human watches, redirects or walks away. Most of them come down to one constraint: the agent’s context window fills fast,…
-
Agentic AI Subagents: One Agent Cannot Check Itself
What are agentic AI subagents? Agentic AI subagents are separate agents that a main agent hands work to. Each one runs in its own context window, with its own tools and its own permissions. So the main agent never sees the mess, only the result. And one of those subagents can be the one that…
-
What Is an Agentic AI Circuit Breaker? A Rehearsal, Not a Button
An agentic AI circuit breaker is the rehearsed ability to stop an agent mid-run and put every system it touched back the way it was, in a time you have measured. Two halves. The stop halts the loop before the next tool call. The rollback undoes the calls that already landed. A button that only…
-
What Is an Agentic AI Autonomy Tier? A Band, Not a Grant
An agentic AI autonomy tier is a named band of actions an agent may take without a human in the loop, with a hard ceiling on what it may not do at that band, and a rule for how it moves up or down. The tier is not a switch. It is a range with…
-
Agentic AI SDLC: From Prompted to Autonomous, Step-by-Step Guide
If you run software programs for a living, you already know the shape of what is coming: a maturity ladder, gates between rungs, a sign-off before the next rung is funded. That instinct is right. What changes with agentic AI is what the gates are made of. A traditional gate was a meeting, a sign-off,…
-
Agentic AI Model Never Decides the Outcome. The Controls Do.
The best agentic AI case studies I have published in the last year share one finding, and it is not about models. Across ten posts covering more than sixty enterprise agentic AI deployments, not one outcome, good or bad, was decided by which model the company picked. Every one was decided by what the agent…
-
Anthropic’s AI-Native SDLC Playbook, Mapped to My Agentic AI Playbook
The AI-native SDLC is a rebuilt software development lifecycle in which AI is embedded at every stage — plan, design, build, test, deploy, maintain — and each stage ends by committing a version-controlled artifact that the next stage reads. That chain of artifacts, from intent.md through the spec, the plan, the diff and the incident…
-
How Fortune 1000 Companies Organize for Agentic AI
Fortune 1000 companies organize for agentic AI in one of three ways: a centralized AI Center of Excellence, fully embedded teams inside business units, or a hub-and-spoke hybrid that combines both. The published evidence now points clearly at the hybrid. IBM’s study of more than 600 Chief AI Officers across 22 countries found that organizations…
-
27 Key Agentic AI Terms, What They Mean and Where They Came From
This is a glossary of 27 key agentic AI terms — what each one means, where it came from, and one authoritative source to read for each. The vocabulary of agentic AI has moved faster than any enterprise glossary I have seen in twenty years of consulting. Terms that were lab jargon eighteen months ago…
-
Agentic AI Model Harness: The Layer That Decides Agent Outcomes
What Is an AI Model Harness? An AI model harness is everything wrapped around a raw language model that turns it into a working system — the system prompts, tool interfaces, context and memory management, the agent loop, orchestration, guardrails, permissions, sandboxing, and observability. The model weights are not the harness. And in 2026, the…
-
What Agentic AI Actually Costs: The Complete P&L Breakdown
The cost of agentic AI is not the price of a model call. It is every token, tool call, retry, verification pass, escalation, and minute of human review it takes to finish one completed unit of work — and in the twelve months ending December 2025, the price per million tokens fell by roughly half…
-
The Coming CxO Power Shakeout Over Agentic AI
A power shakeout is coming in the C-suite over agentic AI: within the next two to three years, CFOs, CMOs, and COOs will move from funding agentic AI deployments to directly managing them, and CIOs and CTOs who keep treating these programs as technology initiatives will find themselves running the plumbing while someone else runs…
-
Cost per Completed Task: Agentic AI’s Unit of Account
Cost per completed task is the total spend an agentic AI system incurs to finish one unit of work — every token, tool call, retry, escalation, and minute of human review, added up across the whole loop. Not the price of one model call. The whole loop. This is the thesis I laid out on…
-
Two Agentic AI Deployments, One Contained, One Catastrophic
The Difference Was Operational Risk Management Summary: Of the two 2025 agentic AI coding deployments compared here, one sits at Stage 3, Governed, on my Agentic AI Roadmap, and one operated at Stage 2, Piloted, while being marketed as Stage 5, Autonomous. The gap between them was not model capability. It was whether the agent’s…
-
Agentic AI Autonomy Trap: Healthcare vs Fintech Case Study
Two Agentic AI Case Studies – One Success, One Reversal And What Separates Them I have spent this year documenting agentic AI wins. This edition is different. I want to put a genuine success story next to a genuine reversal, side by side, because the gap between them is not about the technology. It is…
-
Agentic AI’s Autonomy Trap – A Case Study
I have spent this year writing about agentic AI wins across banking, healthcare, and manufacturing. A European fintech company is the case I keep coming back to for a different reason. It is the clearest public record of a company that pushed straight past Governed toward a claimed Autonomous, skipped the Assured stage of my…
-
The Agentic AI Roadmap Stage Behind Tesla Robotaxi Headline
Tesla turned on unsupervised Robotaxi service in Miami. No safety monitor in the seat, live in a geofenced zone covering West Miami, Doral, and Coral Gables. Florida is now the first state outside Texas to get it, and Miami is the fourth city in the network after Austin, Dallas, and Houston. I read that headline…
-
5 Stages of the Agentic AI Roadmap and Tesla FSD
Everyone wants AI agents that can work on their own. The hard part is knowing what each step toward that looks like, and being honest about which step you are on today. I built the Agentic AI Roadmap to make that clear. It has five stages, from Prompted to Autonomous. You reach each stage by…
-
11 Claude Agentic AI Case Studies Analyzed and Mapped
After going through AWS’s case study library, I went back to Claude’s own customer stories page and did the same exercise. If you are interested in reading Claude agentic AI case studies, this resource is especially useful. It is a different kind of library. AWS shows you everything from security automation to lunar hardware. Claude’s page…
-
11 AWS Agentic AI Case Studies Analyzed and Mapped
I spent this weekend going through the AWS Solutions Library and its customer success stories, hunting for the deployments that have actually shipped. I was specifically interested in AWS agentic AI case studies—not the polished demos, but the ones running in production with real numbers attached to them. Eleven of them are worth your time. I…
-
50 Agentic AI Enterprise Case Studies, Mapped to the Roadmap
This post has one objective: to collate the enterprise agentic AI deployments worth knowing about, analyze what each one actually achieved, and classify it against the Agentic AI Roadmap I published earlier. Most agentic AI conversations stall at the pilot. The more useful question is what happens when an agent reaches production and has to deliver…
-
The Agentic AI Roadmap: A 5-Level Maturity Model From Prompted to Autonomous
The Agentic AI Roadmap is a five-level maturity model for enterprise agentic AI. It runs from Level 1, Prompted, where people use chatbots one at a time, to Level 5, Autonomous, where agents run production work under provable control. Each level is earned by institutionalizing the one below it, and each comes with one governance…
-
Agentic AI Assurance Is Now a Market, for REAL
Something new is showing up in enterprise AI budgets. Not agents. The audit of agents. A market is forming around agent assurance. Independent third-party review. Runtime monitoring. Even guardian agents whose only job is to watch other agents and rein them in when they drift. A recent Fact.MR market study projects this assurance market growing about 44…
-
AI Agents Getting Verifiable Identity: What It Means for Your Enterprise
This third post on Enterprise AI focuses on AI Agents Getting Verifiable Identity and what it means for Enterprise AI. The thread so far: “Intent in, outcomes out. If you cannot draw a straight line from a token spent to an outcome delivered, you are not running an agent. You are running a meter.” — Agentic…
-
Agent Identity Is the Control Plane. Most Enterprises Skip It.
Who Was That? The Agent Identity Question Nobody Can Answer by Dr. Harish Kotadia When an agent acts inside your systems, the first question a regulator, an auditor, or an incident responder will ask is simple. Who was that? For most enterprises, there is no good answer. In a 2026 survey the Cloud Security Alliance…
-
What is Agentic AI? Definition of Agentic AI
A Definition of Agentic AI — From the Practitioner’s Viewpoint “Agentic AI is the next enterprise workload abstraction: a governed, goal-driven software layer in which LLM-powered agents — equipped with memory, tools, and orchestration protocols — autonomously plan and execute multi-step business processes across your cloud, data, and application estate, the way containers once abstracted…
-
I Read 50 Definitions of Agentic AI. Here’s the One the Builders Actually Need
Ask ten experts “What is Agentic AI?” and you will get twelve answers. I know, because I went looking. Over the past few weeks, I systematically reviewed the top 50 definitions of Agentic AI published across peer-reviewed scholarly journals (IEEE Access, Springer’s Artificial Intelligence Review, MDPI Future Internet, F1000Research), the leading industry research firms (Gartner,…
-
Agentic AI: The hidden cost of NOT doing it RIGHT
Most enterprises are still running linear LLM chains and calling it “AI transformation.” Here’s what that’s actually costing you: Fact 1: Organizations running agentic workflows report up to 35% gains in operational efficiency and materially faster task execution. Your competitors aren’t waiting. Fact 2: Gartner says 40% of enterprise apps will embed task-specific AI agents…