March Product Updates: Call Center AI

This month’s updates are designed to make multilingual customer conversations clearer, faster, and more consistent. From improving understanding across accents and expanding language support in translated calls to bringing more structure and flexibility to QA workflows, Call Center AI helps teams work more efficiently and deliver better customer experiences. Let’s take a closer look!

Krisp Call Center AI updates

• Customer Accent Conversion for improved understanding: Customer Accent Conversion helps agents better understand customers during live calls by softening incoming accents in real time. It dynamically adapts to accent variability, while preserving the customer’s natural voice and tone. It works alongside Customer Noise Cancellation to keep incoming audio clear. Agents can enable it anytime from the Krisp app or Widget. Reach out to your Account Executive to learn more.

• Vocabulary suggestions by agents to improve translation accuracy: Agents can now suggest terms directly from the Krisp Widget during live calls. Once reviewed and approved by admins, submissions are added to shared vocabularies to continuously improve translation accuracy across the team.

• Save time with Quick Phrases for Voice Translation: Agents can now select from admin-created phrases, customize them within the Krisp Widget and play them in the customer's language in voice during translated calls. This saves time when communicating repeated disclosures or standard texts and ensures a consistent and accurate delivery.

• Expanded language support and improved voices in Voice Translation: French (Canadian) and Arabic (Egyptian) are now available for synchronous translation across any language pair within Krisp’s 60+ supported languages, on either the agent or customer side. We’ve also improved translated voice quality for Arabic, Dutch, and Portuguese, delivering more natural-sounding accents and clearer phone number pronunciation for smoother, more reliable multilingual conversations.

• QA scorecards for more consistent call evaluations: Supervisors can now evaluate calls using structured scorecards with scores, justifications, and breakdowns in one place inside admin portal. Scoring rules and thresholds are customizable per team, making it easier to standardize evaluations and coach consistently. To enable QA scorecards for your team, contact your Account Executive.

• Manual score editing for more accurate performance data: When an AI-generated call score doesn't reflect the full picture, supervisors and QAs can now adjust it directly on the Call Details page in the admin portal. Every edit requires a note, changes are tracked, and edited scores are visually flagged. Any metrics derived from that score update automatically, keeping performance data accurate.