Agent design patterns

Eight ways agent products lose trust

The failures that recur across agent products. Each one names what it costs and the interaction pattern that addresses it.

What it looks likeWhat it costsInsteadAddressed by
Confirm only at the end
The agent works for twenty minutes and then asks whether it got it right
One error in step three means redoing all twelvePlace checkpoints deliberately, weighted earlyScheduled confirmation
Transparency theatre
A wall of tool calls labelled reasoning
Overload reads as opacity and nobody checks anythingSummary by default, detail on requestAltitude control
One autonomy switch for everything
A single on or off covering every task
One bad experience turns the whole feature offOne dial per task typeAutonomy dial
Silent irreversible action
It already happened, with no preview and no undo
Trust rarely returns after thisPreview, undo, and an honest expiryIntent preview
Notification flood
Several agents each interrupting on their own schedule
People stop reading, including the one that matteredOne attention budget across every agentInterruption budget
Chat as the container for everything
Parallel and overnight work forced through one transcript
Nobody can say what is happening right nowA decision queue and a run boardAgent inbox
Anonymous fleets
Agents with no name, owner or remit
Nobody can say who authorized whatOwner, scope, spend cap, review dateAgent registry
Confident guessing
Ambiguity resolved quietly into something plausible
The error arrives finished and formattedThree ways out: notify, ask, reviewEscalation triad

All agent design patterns

Review an agent feature with us30 minor send a message