A Building Agentic AI

Building Agentic AI

Build agentic AI that works in production.

Practical field notes, articles, and hands-on help for engineers, architects, and technology leaders building agentic AI systems in real organizations.

Start Here See how I can help

Muhammad Arbab
Written by Muhammad Arbab. 20 years in enterprise systems, 14 of them shipping AI inside Fortune 100 organizations. Author of two books on designing agentic AI systems.

Try it

Watch the agent work, and book a real call.

A scheduling assistant built as an agent. Its reasoning and tool calls appear in the transcript as it runs. Chat with it and book a 30-minute intro on Muhammad's calendar.

Open the assistant

The point

An agent is software that uses a language model to decide what to do next, takes that action, looks at what came back, and decides what to do after that. Everything else is plumbing.

from What counts as agentic AI, and what does not

  • Jun 07, 2026

    Why Voice AI Agents Are Harder Than Chatbots

    A working chatbot rarely survives the jump to a phone line. Why voice agents are harder: latency as a hard budget, barge-in, ASR errors, silence that means something, emotion, and real-time handoff.

    Enterprise
  • Jun 07, 2026

    What to Automate First: A Leader's Framework for Agentic AI in 2026

    The question that stalls most agentic AI programs is not "can we build an agent." It is "which process first." A leader's framework for choosing the first project: four scoring criteria, one hard gate (blast radius), a 2x2 you can draw on a whiteboard, and a worked example scoring three real processes.

    Enterprise
  • Jun 06, 2026

    When to Use an Agent Framework vs Build It Yourself (LangGraph, CrewAI, and More)

    The honest answer is not "always" or "never." Build the raw loop yourself until you hit a specific named pain, then adopt a framework to solve that pain, not on principle. A decision diagram, a map of the frameworks worth knowing (LangGraph, CrewAI, AG2, LlamaIndex, OpenAI Agents SDK, Pydantic AI, plus the AWS, Microsoft, and Google platforms) and what each is good at, and the signals that tell you which side of the line you are on.

    Beginner
  • May 26, 2026

    Agent Failure Modes Interviewers Probe: Tool Misuse, Loops, Prompt Injection

    Senior interviewers do not ask how agents work. They ask how agents break. The seven failure modes that decide most agentic AI system-design rounds in 2026, the follow-up questions interviewers actually use, the structured answer template, and the two failures that get candidates rejected when missed.

    Enterprise

Newsletter

Subscribe, get Volume I free.

Get Understanding Agentic AI Systems free when you subscribe, plus a short note when new articles ship. No spam, no drip sequences, unsubscribe in one click.

Powered by Kit · GDPR-friendly