Intercom Adds Custom Pronunciation Rules and 11 New Voices to Fin Voice Agent.

Intercom has rolled out three significant upgrades to Fin Voice, its AI-powered phone support agent, addressing some of the most common friction points in automated voice interactions.

The updates, announced on LinkedIn by the company's product team, include custom pronunciation rules, an expanded library of 11 natural-sounding voices, and enhanced background noise reduction. The changes aim to make Fin Voice sound less robotic and more reliable when handling real-world customer calls.

The standout feature is pronunciation rules, which let support teams define exactly how Fin Voice should say brand names, product names, and industry-specific terminology. This addresses a persistent problem with AI voice agents: mispronouncing unique or technical terms can immediately undermine customer trust.

For companies with unconventional product names or abbreviations that standard text-to-speech systems struggle with, this granular control could be the difference between a voice agent that sounds competent and one that confuses callers. Teams can now ensure Fin pronounces "SaaS" as "sass" rather than spelling it out, or handles proprietary product names the way customers actually say them.

Intercom also expanded Fin Voice's vocal range with 11 new voices. The company says this gives teams more flexibility to match their brand's tone, whether that means sounding warm and approachable or crisp and professional. Voice selection has become table stakes for AI agents as businesses look to maintain consistent brand identity across all customer touchpoints.

The third upgrade tackles a practical challenge: customers call from noisy environments. Fin Voice now includes automatic background noise reduction, designed to filter out ambient sound and keep conversations intelligible. Whether someone is calling from a busy office, a coffee shop, or outdoors, the system attempts to isolate their voice and maintain natural conversation flow.

This isn't just about audio quality. When voice agents struggle to understand callers due to background noise, conversations break down quickly, forcing escalation to human agents and defeating the purpose of automation.

The timing of these updates aligns with Intercom's broader push into AI automation. The company has been steadily expanding Fin's capabilities, recently launching a complete observability suite that helps teams monitor and improve AI agent performance.

Intercom demonstrated the new features in a product walkthrough featuring team members Brian and Alihan, though the company has not yet published detailed documentation on implementation. The updates appear to be rolling out to existing Fin Voice customers as part of the platform's standard offering.

For businesses evaluating voice AI agents, pronunciation control and voice customization are increasingly non-negotiable. As these systems handle more complex support scenarios, the gap between "technically functional" and "actually trustworthy" often comes down to these kinds of refinement details.