Bob Michaels/ai
Published by Bob Michaels31 pieces · 19 essays · 6 episodes · 6 shorts

Library.

One argument, three formats. Read it, listen to it, or watch the sixty-second version.

There is no CMS behind this page. Every essay and every episode is a typed entry in a registry and a file in the repository. The index, the sitemap, and both feeds read that one source, so publishing is a commit and nothing can drift out of sync.

Follow the work

EssaysRSSPodcastFeedShortsYouTube ↗

Newest first.

Episode 06

August 5, 2026

Built with AI

13:43 listen

What Happened to WordPress

WordPress started life as blogging software in 2003 and grew sideways into the thing that runs about four out of every ten websites. For the first time in fifteen years, that number is going down. Bob Michaels, building the web since 1994, explains what a content management system was for, the hidden bill that came with it in add-ons and themes, the growing pile of doors a modern site has to keep open from sitemaps to agent to agent, and why the problem it was all built to solve is no longer a problem.

Listen to the episode →

Episode 05

July 30, 2026

Built with AI

7:26 listen

Who Signs Off When the AI Acts?

Human in the loop means a person is present. Human in control means a person can inspect the evidence, stop the action, and own the decision. Bob Michaels walks through the approval map to draw before you automate anything, and the seven actions an AI agent should never take without a named human owner.

Listen to the episode →

Essay

July 30, 2026

Built with AI

5 min read

The CMS Era Is Sunsetting

WordPress peaked at 43.6% of all websites in 2025 and posted its first decline since W3Techs began tracking in 2011. Bob Michaels, building the web since 1994, argues the CMS era is sunsetting: websites are becoming software again, built with AI, readable by AI, and transacting with AI.

Read the essay →

Essay

July 28, 2026

Built with AI

10 min read

Human in Control: Draw the Approval Map Before the Workflow

Human in the loop usually means a person is present. Human in control means a person can inspect the evidence, stop the action, and own the decision. Here is the approval map to draw before you automate: action classes, named owners, evidence packets, four decision states, and the thresholds that keep it from becoming bureaucracy.

Read the essay →

Episode 04

July 23, 2026

Built with AI

8:03 listen

Can You Trust What the AI Just Told You?

An AI system you can trust at work is one that cites real evidence, admits uncertainty, or declines to answer. Bob Michaels walks through two weeks of essays on what grounded actually means, where your company knowledge should live, why betting everything on one model is a risk you chose, and how to test an AI agent before it touches production.

Listen to the episode →

Essay

July 23, 2026

Built with AI

9 min read

How to Evaluate an Agent Before It Touches Production

An AI agent is a stateful business process with tools and permissions, and it should be evaluated like one. Here is the four-class test matrix (normal, edge, adversarial, forbidden), the trace record, the two scores that must stay separate, and the release gates that decide production.

Read the essay →

Essay

July 21, 2026

Built with AI

9 min read

One Model Is Not an AI Strategy

A multi-model strategy is a stable task contract with portable evidence, and the vendors' own documentation explains why you need one: models retire on the provider's clock, notice windows run from two weeks to six months, and compatible APIs cover the request shape, never the behavior. Here is the dependency map and the exit test.

Read the essay →

Essay

July 14, 2026

Built with AI

9 min read

Grounded Means Cited or Declined

Retrieval is a component. Grounding is a contract: for consequential claims, the system cites approved evidence, states uncertainty, or declines, and logs which one happened. Here is the contract, the standard behind it, and the test that tells you whether a system honors it.

Read the essay →

Episode 03

July 9, 2026

Readable by AI

7:53 listen

Who Does AI Say Your Competitors Are?

When a buyer asks AI who to hire, the answer names a winner and a field of competitors, and that field is discovered, not assumed. Bob Michaels walks through two weeks of essays on the head to head, the only AI visibility metric shaped like the business question, why the standard content playbook cannot answer it, and the loop that moves the recommendation.

Listen to the episode →

Essay

July 9, 2026

Readable by AI

8 min read

How You Win the AI Head-to-Head

Losing the AI comparison is a finding, and findings have a method. Here is the loop I run to move the recommendation: query on a locked list, decode why the winner wins, engineer that evidence on your domain with one consolidated message, and verify by retesting. With the measured results, honestly framed.

Read the essay →

Essay

July 2, 2026

Readable by AI

8 min read

Stop Topic Chasing. Start Winning the Head-to-Head.

Topic chasing is the keyword-and-coverage playbook, and it measures inputs. The head-to-head measures who AI actually recommends. Here is the difference, the evidence for both halves, and how to keep your content program while changing its finish line.

Read the essay →

Episode 02

June 25, 2026

Readable by AI

7:33 listen

Does Your Website Tell the Truth?

Machines read everything you have ever published, not just the homepage you keep polished. Bob Michaels walks through two weeks of essays on the website truth audit, how to move a site without erasing what search and AI already know about you, what machine readable actually means, and why accessibility is now an architecture requirement.

Listen to the episode →

Essay

June 25, 2026

Built with AI

7 min read

Accessibility Is an AI Architecture Requirement

AI interfaces stream text, update live, and fail in new ways, and every one of those behaviors has an accessibility contract. Here is how WCAG maps onto chat and agent interfaces, and why it belongs in architecture, not remediation.

Read the essay →

Episode 01

June 11, 2026

Readable by AI

7:49 listen

Does AI Even Know You Exist?

Your customers are asking ChatGPT who to hire, and most companies have no idea what it says about them. Bob Michaels walks through the four essays he published over the last two weeks, covering the five things AI visibility actually measures, why a citation is not proof, how to ask the questions so the answers count, and how to turn what you find into a thirty day plan.

Listen to the episode →

Essay

May 28, 2026

Built with AI

9 min read

The Seven Actions an AI Agent Should Never Take Without a Named Owner

Agent permissions are being granted by integration scope instead of business consequence. Seven action classes need a named human owner no matter how automated execution gets: external communication, publication, financial commitment, access changes, destructive data actions, production deployment, and legal representation. Here is the list, with the bounded exceptions that keep it fast.

Read the essay →

Shorts.

Subscribe on YouTube ↗
What is Web 3

Platform shifts

What is Web 3

What is a Vector Store

Production AI

What is a Vector Store

Content is Code

AI visibility

Content is Code

The New WWW

The post-CMS web

The New WWW

Content Generation

AI visibility

Content Generation

The Bots Are Coming

The post-CMS web

The Bots Are Coming

How this page works

No database, no admin panel, no plugins.

Every piece here is a typed entry in a registry and a file in the repository. The list you are reading, the sitemap, the RSS feed, and the podcast feed all derive from that one source. I build client sites the same way.

See the method →Talk to me directly