Collateral: Augmented Recognition
Parlance HIPAA-compliant AI voice assistants transform healthcare switchboard and call center operations by automating workflows at the digital front door. For over 25 years, Parlance has been a leader in the healthcare IT space. Our conversational AI solutions for health systems accelerate patient access to care for thousands of hospitals and clinics across the United States. Leading healthcare organizations, enterprises, and higher education institutions rely on Parlance, as a leading provider of IVR and IVA solutions.Voice-first AI that delivers measurable ROI in 30 days
Summary
Healthcare IT leaders face a critical challenge: traditional voice AI systems fail to understand 15-20% of caller responses, resulting in lost patient connections and operator escalations. Augmented recognition technology addresses this gap through real-time reprocessing of unclear utterances. The Parlance HIPAA-compliant solution combines on-premises security with cloud-based speech-to-text capabilities, significantly improving call completion rates without additional staffing. Parlance delivers measurable ROI within 30 days by handling challenging audio conditions — including noisy environments, cross-talk, and verbose responses. With SOC2 compliance, encrypted data transmission, and isolated cloud instances processing only low-confidence utterances (3-5 seconds), augmented recognition ensures operational efficiency while maintaining stringent healthcare data protection standards for patient access excellence.
What is Parlance Augmented Recognition?
Parlance drives patient access excellence — Healthcare organizations lose patient connections when traditional voice AI fails to understand callers. Parlance is a fully-managed HIPAA-compliant solution that uses augmented recognition to address this critical gap by reprocessing unclear responses in real-time, significantly improving call completion rates and reducing costly operator escalations.
For IT leaders focused on operational efficiency, augmented recognition delivers measurable improvements in consumer satisfaction and contact center performance, ensuring more callers reach their destination on the first attempt without adding staff resources.
Recognition can be improved for several audio conditions, including:
- Responses that include more information than expected
- Noisy environments with cross-talk or background music
How Does Augmented Recognition Work?
Augmented recognition adds a cloud-based speech-to-text (STT) service to your on-premises solution. This ‘hybrid cloud’ architecture combines the best of both worlds: keep your most critical systems on-premises while leveraging high value and secure cloud services.
Why Parlance?
As a fully-managed service, Parlance regularly evaluates and adopts emerging technologies to ensure that customers benefit from recent innovations. The Parlance platform provides real-time transcription, best-in-class AI-accuracy, and SOC2 compliance.
Parlance augmented recognition consists of:
- A secure instance that completely isolates all customer data in the cloud
- Protected data in transit: encryption + authentication
- Limited data use: Only low-confidence responses are sent
- HIPAA compliance: Responses are used for name-matching only and not shared
Protecting Customer Data
Parlance provides only the recording of the caller’s utterance only when responding to a question from the application (where typical utterances are 3-5 seconds in duration). When using speech-to-text services, Parlance maintains the security of customer data with the following practices:
- Uniquely identifiable information is not provided to speech-to-text services.
- All data, including audio, is encrypted in transit. To access Parlance’s cloud-based APIs, Parlance uses SSL connections using Transport Layer Socket (TLS).
- No data, including audio, is stored by our speech-to-text service. Data provisioned to the Parlance cloud is maintained for runtime use only. When the runtime environment changes in response to directory or other related updates, data associated with the prior runtime state is discarded.
- Parlance recommends that provisioned external access be well-constrained. To support Parlance augmented recognition, external access can be limited to outbound-initiated requests to specific URLs. Please contact your solutions engineer for the URLs.
