The analyst could be impressed.
The buyer could be charmed.
The machine can be neither.
All three were checking for the same thing.
Coherence was always the standard. The machine is just the first reader honest enough to enforce it.
https://t.co/U3vMaEmYdv
If you build demos, read yours the way the machine will, before it does. Demo Intelligence: https://t.co/724HkelVEC Video 1 of 3. The Root Cause, Issue 07: The Literal Reader. https://t.co/SS96eZy4Dh
A person fills your gaps. The machine doesn't. It reads literally, keeps what it can verify, and drops the rest. "Enterprise ready." "Trusted by teams like yours." Gone. "Syncs with Salesforce and SAP." "Returns results in 1.2s." Those it keeps, because they're checkable.
Gartner has a name for the practice. Agentwashing. Labeling AI assistants as agents when they do not operate independently. On the demo pod floor it shows up differently. Agentwashing is what vendors do. Agent-shaped is what they ship.
Most of what is being sold as agentic AI in 2026 cannot survive thirty seconds on a real demo pod floor.
For twelve years I sat on the other side of those demos.
Field Note 06: Demo is the Diagnosis.
Watching this quarter's agent platform pitches.
Every vendor names "agents" by what they appear to replace.
Assistant. Copilot. Worker.
Same mistake auto did with the horseless carriage.
https://t.co/BeioZnmfC7
The Horseless Carriage Problem
Most enterprise AI demos are selling a Category 3 story.
What shows up on screen is Category 1.
Buyers feel that gap. They just can't name it.
https://t.co/rioGJkmSbJ
By 2024 the most credible thing an AI marketer could do was lead with what AI had gotten wrong.
Not as spin. As the actual opening.
Buyers had been through one round of broken promises. They knew what the pitch sounded like.
Three phases. What each one actually cost.