Tag: agentic ai
-
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 Code Review: Intent, Not Diff
What is agentic AI code review? Agentic AI code review is how a team checks code that an agent wrote before it ships. The human stops reading every line and reviews the intent, the evals and the scope of the change, while review agents scan the diff first. One thing stays fixed: a named person…
-
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 Agent-to-Agent: A2A Settles Handoff, Not Trust
What is agentic AI agent-to-agent trust? Agentic AI agent-to-agent trust is the decision to let one agent act on what another agent says. The A2A protocol settles how two agents find each other and hand a task across. But it does not settle whether I should believe the agent on the other end. So a…
-
Agentic AI Architecture Map
Every post on agenticaiarch.com: by the six-layer architecture that decides outcomes, and by the Five-Stage Roadmap an agent program moves through. Click any post to open it: Six-Layer Agentic AI Architecture Pillar: What Is Agentic AI Architecture? Pillar & Foundations The map itself, the vocabulary, and the definitions everything else hangs on What Is Agentic…
-
Agentic AI Non-Determinism: Repeat the Boundary, Not the Route
What is agentic AI non-determinism? Agentic AI non-determinism is the fact that the same agent, given the same task twice, takes a different route each time. The prompt is the same, and so is the model. But the steps, the tool calls, and sometimes the answer are not. So a test that expects the same…
-
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 MCP Server Governance: Vet the Server, Not the Call
What is agentic AI MCP server governance? Agentic AI MCP server governance is the set of controls that decides which MCP servers an agent may connect to, who decides that, and what a user can change. It works at the server level, not the tool-call level. So the question moves from “was this call safe?”…
-
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…
-
Agentic AI Egress Allowlist: Stop It Calling Out
What is an agentic AI egress allowlist? An agentic AI egress allowlist is a short list of outside hosts an agent may call. The network blocks every other host by default. So the list sits outside the agent, and no prompt can talk its way past it. You cannot stop an agent thinking. You can…
-
Agentic AI Human-in-the-Loop Gate: Gate What Matters
What is an agentic AI human-in-the-loop gate? An agentic AI human-in-the-loop gate is a checkpoint in the harness. It pauses one class of agent action until a named person approves it. The gate sits on the action, not on the prompt. So it fires only when the agent tries to do something that matters. New…
-
Agentic AI Branch Protection: The Agent That Wrote It Cannot Approve It
What is agentic AI branch protection? Agentic AI branch protection is a rule on the repository that stops any pull request from merging until someone other than its author approves it. The agent that wrote the change cannot approve it. That rule lives in GitHub, not in the agent, so the agent cannot talk its…
-
Agentic AI Credential Scoping: One Key, Not Yours
What is Agentic AI credential scoping? Agentic AI credential scoping is the practice of giving an agent its own short-lived, narrow credential instead of the developer’s. The agent gets one key, cut to one job, that expires. It never gets yours. Yesterday’s post was about the rule the developer cannot edit. This one is about…
-
What Is an Agentic AI Managed Setting? A Rule the Developer Cannot Edit
Agentic AI managed settings are the permission rules an administrator sets above the developer. No project file, local file, or command-line flag can loosen them. That is the whole idea. The rest of this post is the how. Last week I wrote about the circuit breaker and prompt injection. Both of those assume the rules…
-
Keeping Agentic AI Systems Safe From Prompt Injection
Agentic AI prompt injection is what happens when an agent reads content that contains instructions and then follows them as if you had given them. The content can be a web page, an email, a PDF, a support ticket, or the result of a tool call. You asked the agent to do one job. Then…
-
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 Agentic AI Context Caching? Stop Resending the Codebase
What Is Agentic AI Context Caching? Agentic AI context caching is the runtime control that lets an agent pay once to process a large, stable block of context. On every later turn, the agent reuses the provider’s stored computation at about a tenth of the price. The block is usually the system prompt, the tool…
-
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…
-
Enterprise Agentic AI roadmap for 2027
An enterprise agentic AI roadmap for 2027 is a governance plan with an architecture inside it. For a Fortune 1000 company it is the year the weakest production agent climbs from Piloted to Governed, and the best one earns Assured, one exit gate at a time. This is the fall 2026 edition, the one about…
-
What Is Agentic AI Context Compaction? Decide What Goes
Agentic AI context compaction is the step in an agent’s runtime loop where the harness takes a conversation that is nearing the context window limit, has the model write a summary of it, and restarts the loop with that summary in place of the original history. Everything before the summary is gone from the model’s…
-
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…
-
Agile in Agentic SDLC? Keep the Ceremony. Retire the Artifact.
An agentic SDLC is a software development lifecycle in which AI agents complete whole units of work under enforced boundaries, and it does not retire agile. It retires agile’s artifacts. Every ceremony you run still has a reason to exist. What each one produces has to stop being a document a human interprets and start…
-
What Is Agentic AI Architecture? The Six Layers That Actually Decide Outcomes
Agentic AI architecture is the design of the system around the model: the layers that decide what an agent is asked to do, what it may reach, what it is forbidden to do, how far it may act without a person, and how any of that is proved afterwards. The model is a component in…
-
What Is intent.md? Treat It as a Regulated Record from Day One
In practice the intent stage produces one artifact, intent.md, which I treat as a regulated record from day one. An intent.md is a short, version-controlled Markdown file that records the ask, the reason behind it, and the constraints it has to respect — committed before anyone designs or builds anything. It is the file Claude Code…
-
What Is an AI-Native SDLC? Intent Becomes the Unit of Work
An AI-native SDLC is a software development lifecycle in which the unit of work is committed intent rather than an assigned instruction, and each stage ends by writing a version-controlled artifact that the next stage reads. Anthropic published the clearest description of one on August 21, 2026 in The AI-Native SDLC playbook, and has since…
-
10 Facts You Wish You Knew Before Building Agentic AI Solution
Most agentic AI projects fail on facts that were knowable before anyone wrote code. Here are the ten I most wish every architect and CIO knew going in. None of them are about how smart the model is. Every one is a property of the model you rent; every fix is a property of the…
-
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…
-
How the CxO Shakeout Rewires IT Services Firms
The CxO power shakeout over agentic AI will force IT services and consulting companies to rebuild their sales, marketing, and delivery model around business executives — CFOs, COOs, CMOs — instead of the CIO and CTO organizations they have sold to for thirty years, and to staff those accounts with deep domain experts rather than…
-
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…
-
Why I Wrote This Book on Agentic AI
Over the past eight months, I have had the opportunity to speak with senior technology professionals at Fortune 500 companies while evangelizing agentic AI solutions. What I realized surprised me: many of them were not doing agentic AI right. Not because of lack of talent or budget, but because they lacked a clear definition of…
-
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…
-
Agentic AI in Healthcare: The Prior Authorization Case Study
For as long as I have been building agentic AI systems, I have heard a version of the same objection when the conversation turns to healthcare. Protected health information cannot touch an AI agent, they say. The liability is too high, the compliance burden too strict, the margin for a confident wrong answer too thin.…
-
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…
-
From Siebel to Agentic AI: How Enterprise Data Finally Got Free
I have spent over twenty years installing software so that business people could look at their own customer data. I mean that literally. Some of my earliest projects involved putting a CRM client on individual desktops, one machine at a time, before a salesperson could see a single account they already owned. That should sound…
-
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: You’re Not Running an Agent. You’re Running a Meter.
Why Agentic AI Cost Is an Architecture Problem By Dr. Harish Kotadia, Ph.D. TechCrunch reported that one company received a roughly $500 million bill in a single month, after it forgot to set a usage cap. The token price went down. The bills went up. That is not a pricing problem. It is an architecture…
-
Six Lessons Learnt: Claude Implementation
Pitfalls and Lessons learnt: By Dr. Harish Kotadia, Ph.D. Our team delivered an agentic auto loan origination program on Claude. The biggest risk on the plan was never the model. When reflecting on this project, the team discovered unique challenges specifically relevant to managing a Claude auto loan system. It was scope. We had a…
-
Agentic AI for the Enterprise: Practitioner’s Field Notes
Agentic AI for the Enterprise: One Practitioner’s Field Notes By Dr. Harish Kotadia, Ph.D. I have spent close to two decades building enterprise systems for Fortune 100 clients. I have watched three real shifts in that time, and I am convinced the one happening now is the largest of them. Over the past few weeks…
-
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…
-
The Question Every IT Services Firm Is About to Be Asked
A few months ago I was in a working session with a client’s procurement team, and one of them asked a question I haven’t been able to shake since. We were reviewing a managed services renewal, and she said, almost casually: “If your delivery teams are using AI now, why is the price the same…
-
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: 50 Authoritative Definitions — The Complete Citation Compendium (2026)
What exactly is Agentic AI? Ask ten experts and you’ll get ten answers — and that’s not a bug, it’s the most important insight in this space right now. Multiple credible sources, from MIT Sloan Management Review to Gartner, explicitly acknowledge there is no single universally agreed-upon definition. This compendium aggregates 50 authoritative definitions drawn…
-
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…
-
Agentic AI: Why cloud platform choice matters more than you think
Which cloud platform to choose for Agentic AI? This isn’t purely an AI question. It’s an infrastructure question — and both Amazon and Google have built managed services specifically designed to operationalize this loop. That matters a great deal for risk, compliance and governance teams who are trying to figure out how to let AI…
-
CIO guide to Agentic AI
Most Enterprise AI Is Built on the Wrong LayerThe shift from stateless LLMs to agentic AI architecture — and what CIOs should do about it now. Three major technology shifts have reshaped enterprise IT over the past two decades. The current one — the move from stateless LLM calls to agentic AI architectures — is…
-
Agentic AI: It Is Not the Model. It Is the Loop
I have been building IT systems for Fortune 100 clients for two decades. I have watched three major shifts in how enterprises think about technology. The one happening right now is the most significant of the three and most organizations are not ready for it. In this post, we will explore how the Agentic AI Loop…
-
Agentic AI Era: Model is No Longer Product. The Loop Is.
The agentic AI era is here — and Claude by Anthropic is at the center of it.One of the most important concepts to understand in this space is the Agentic AI Loop. What excites me most right now: 🔹 Model Context Protocol (MCP) — standardizing how AI agents talk to tools and data sources, enabling…