Articles

Notes on building agents that businesses can actually trust.

What is changing in agentic AI and customer experience, and what it means if you run support, service or operations. New pieces Monday, Wednesday and Friday.

A man holding an access card up to a card reader mounted on a concrete wall
Agent security • Aug 14, 2026 • 5 min read

Your AI agents have logins. Most of them have no owner.

Security teams can now list every AI agent running in a business. Listing is not control. What matters is what each agent may do, and who decided that.

A metal gate standing open with a padlock still fastened to the frame
Agent security • Aug 12, 2026 • 5 min read

Your AI agent's guardrails live in a file it can edit.

Red teamers took root on an agentic AI through nothing but its chat box. The lesson is not tougher wording, it is putting rules where the agent cannot reach.

An industrial pressure gauge mounted on a stainless steel tank beside a valve
Operations • Aug 10, 2026 • 5 min read

Nothing stops a looping agent. Give it a ceiling.

AWS added rate limits to its AI gateway. An agent that retries has no natural stopping point, so the ceiling has to sit outside it, on infrastructure you run.

A security officer alone in a dim control room facing a wall of surveillance monitors
Architecture • Aug 7, 2026 • 5 min read

Your AI security agent sees one console. You run five.

Microsoft's agentic security works because Microsoft owns the control planes. Almost no business runs one vendor, so the agent that joins them has to be yours.

Rows of metal apartment mailboxes arranged in a symmetrical grid
Platform choice • Aug 5, 2026 • 5 min read

Eleven agent platforms. Which one do you actually own?

A roundup of 11 agentic AI platforms shows how many of them live inside software you already rent. Where the agent runs decides what you keep at renewal.

A team seated at computer workstations monitoring a large digital display in a control room
Operations • Aug 3, 2026 • 5 min read

Your AI pilot works. Who still owns it in month seven?

Enterprise AI spending hit $107 billion and the biggest firms now compete on running systems after go-live. That tells you where projects actually fail.

No articles match that search.