Berk Bayri

Containment rate

The share of customer interactions that stay inside automation without moving to a human agent. A common chatbot KPI that can look healthy while the channel itself weakens.

Containment rate tells you whether an interaction stayed inside automation without moving to a human. It is one of the most common headline metrics for chatbots and virtual agents because it is easy to count and maps neatly to avoided agent minutes.

What it cannot tell you is whether the customer was served well, or whether they will choose automation again. A customer who gives up, abandons the chat, or quietly phones instead may still be counted as "contained".

Reading it responsibly

Containment can look good while the channel gets weaker. Pair it with measures of the next interaction: the return-to-automation rate and repeat-channel economics. Evaluate it inside the session economics of the contact, and remember that sometimes the right move is channel preservation: letting the interaction leave automation early.

Read more in Your chatbot is borrowing from the next interaction.

Used in these essays

Problem Framing & Decision Design13 min read

The safest answer your AI team can give you may be “we should not automate this”

The responsible AI decision is not always a deployment. It is choosing the least autonomous design that improves the outcome without transferring hidden costs or authority.

AI Agents & Systems11 min read

Design systems need an interaction policy layer for AI

Components can keep an AI experience visually on-system while the agent is behaviorally off-policy. Design systems now need reusable rules for what AI may present, decide, execute, defer, expose as uncertain, confirm, reverse and escalate.

AI Strategy & Transformation7 min read

Make the AI vendor demo fail

A polished AI demo proves that a system can succeed under prepared conditions. A buying decision needs different evidence: what happens when the system is wrong, blocked, uncertain or halfway through an action.

AI Governance & Evaluation8 min read

The first AI incident report should be incomplete

OpenAI's new misalignment disclosure framework exposes a useful enterprise design principle: record anomalous AI behavior before the organization has finished explaining it. Otherwise incident systems quietly become filters for what teams already understand.

AI Strategy & Transformation8 min read

Your chatbot is borrowing from the next interaction

AI customer service is usually measured one interaction at a time. But a failed automated interaction can change which channel a customer chooses next time. That makes future adoption part of the economics, not a separate trust metric.