Page 19
12 articlesThe Agent That Could Not Say Wait
Production agents can act, answer, or ask — but cannot say wait. Why the missing primitive forces hesitation into fake tool calls and overconfident commitments, and how to put deliberation back into the protocol.
The Agent That Learned to Hedge Its Way to a Higher Eval Score
An LLM eval score climbing for months while customer satisfaction flatlines is the signature of judge specification gaming. Here is how hedging tics, same-family priors, and missing human calibration combine — and the audit, rotation, and adversarial-slice disciplines that catch it.
The AI Standup Where Yesterday's Status Is a Lie
When agents run overnight and finish three hours after the standup, the round-robin breaks. A queue snapshot read off a dashboard is closer to honest reporting than the Scrum three-questions.
The ChatOps Bot That Mistook Silence for Consent
When a ChatOps bot stops getting replies, the dashboard reads steady-state — but mute, re-asks, and sidecar actions tell the real story. A playbook for instrumenting agents around silence.
The Evidence Locker Your Agent Doesn't Keep
Traces tell you what the agent did. Decision records tell you what the agent had to work with. Most teams shipped only one and will discover the gap during an audit.
The Marketing Page Your RAG Cited as an Engineering Spec
Relevance and authority are different dimensions, and the standard RAG stack collapses them into one score. Here is why polished marketing copy beats your engineering RFC in the vector race, and what to do about it.
The Planner That Treated Every Tool as O(1)
Why agent planners pick correct-but-wildly-expensive tool sequences, and the schema-level changes that make planning cost-aware without retraining the model.
The Pointer Your Agent Mistook for a Value: Reference vs Value in Tool Outputs
Tool outputs that hand back IDs, paths, or URLs ask the agent to dereference them. The model's policy for when to resolve and when to compose-as-if-resolved is implicit, inconsistent, and silently wrong. Type the indirection.
The PR-Bot That Never Sleeps: When Your Reviewers Become the Rate Limiter
AI coding agents now open pull requests faster than humans can read them, turning reviewers into the rate limiter. Risk-tiered auto-merge, review budgets, and AI-on-AI triage are how teams keep throughput honest without rubber-stamping code into production.
The PR Description Your Coding Agent Cannot Write
Agents ship clean PRs with empty descriptions, and async review breaks because the rationale lived in the prompt the harness threw away.
The Prompt Log Is the Product Roadmap You Threw Away
Your AI feature's prompt log is the highest-resolution product discovery signal you have — and the one nobody on your product team is reading. Here's how to mine it for unmet demand.
The Redaction Layer Your Agent Cannot Reason Through
Privacy redaction can preserve classification accuracy while quietly destroying the entity continuity multi-step agents depend on. The fix lives in how placeholders are scoped, not whether they exist.
About Tian Pan
I'm Tian Pan, an engineer-founder focused on agentic engineering — building autonomous AI systems and scaling engineering teams. I write practical guides on system design, technical leadership, and shipping with AI agents. Previously an early engineer at Uber, Brex, and IoTeX.
