What it does
Voice Pattern Mirroring lets an agent quietly observe and adapt to how a specific person communicates over time. Formality level, average sentence length, vocabulary range, preferred greetings, and even punctuation habits get tracked as part of the agent's working memory for that contact.
The adaptation is intentional and bounded: the agent doesn't mimic the customer verbatim, doesn't lose its own voice, and never crosses into impersonation. The goal is to feel native to the conversation, not to perform.
When to use it
- Customer-facing agents that handle conversations across diverse audiences
- Sales follow-up where matching the prospect's energy matters
- Internal coordination across teams that speak in very different registers
When not to use it
- Compliance contexts where every word matters and tone variation would create risk
- Brand voice contexts where the agent must hold a specific tone regardless of the customer
How it works
The skill maintains a per-contact voice profile in the agent's working memory. After every interaction, the profile is updated with new observations. Adaptation happens gradually — large swings in tone trigger a stability check before any change is applied.