Voice Command & Control

Add voice commands to devices, mobile or web applications to elevate user experience.

Work with an expert

Analyze and utilize voice data to offer better products and stay compliant

Understanding voice commands is the core functionality of voice-activated applications such as voice assistants. Voice assistants such as Alexa, Google and Siri increased the adoption of voice control. “Ecobee, decrease the temperature by 1 degree” and “Alexa, turn off the lights” are popular voice commands. The standard approach to processing voice commands has three steps. First, voice input is recorded and then sent to the cloud for processing. After understanding the user intent, the command is sent to the application. Any fickle in the connectivity causes delays, hence poor user experience. In such cases, users conclude that Alexa couldn’t perform the task even if it’s a connectivity issue. Connectivity issues of enterprise applications may hinder employee productivity, resulting in hidden costs. Plus, processing voice data in the cloud is already expensive. Cloud providers such as Amazon and Google cover the cost of their assistants, but not others.

Moving to Picovoice for Numina Group’s Victory Voice solution provided a robust speaker-independent voice recognizer. Picovoice is fast, accurate and supports multiple languages. Both the software tools and technical support services are top-notch. The team is great to work with, they are responsive and accommodating.
Mark Woodworth
Co-founder and VP of R&D, Numina Group

Improve user experience with the most accurate engine

Open-source, open-data NLU benchmark results show Picovoice Rhino outperforms alternative NLU engines, such as Amazon Lex, Google Dialogflow, IBM Watson and Microsoft Luis. Rhino Speech-to-Intent fuses speech-to-text (STT) and natural language understanding (NLU) to give all you need to add voice commands.

Amazon, Google, IBM and Microsoft initially designed their engines for text-based chatbots. This is why they’re also known as chatbots or conversational AI engines.

Encourage user engagement, not hefty cloud bills

Every product owner wants users to interact with its product frequently. However, Natural Language Understanding (NLU) pricing comparison shows that Amazon Lex, Google Dialogflow, IBM Watson and Microsoft LUIS can break the bank if a user interacts with the voice feature heavily. Rhino Speech-to-Intent with unlimited voice interactions lets you focus on increasing engagement.

Let users do what they want when they want, instantly

Bringing voice recognition engines close to voice data eliminates cloud-related problems, such as connectivity, latency and privacy. Therefore users enjoy fast and responsive voice experiences even if they’re in a large warehouse, elevator or countryside. After activating the demo, try the voice enabled coffee-maker demo without an internet connection.


Voice Control

Voice control replaces or complements existing interaction interfaces. In the case of smart speakers, wearables or VR glasses, where physical space is limited, voice can be the only input source. Enterprises add voice control to various applications such as warehouse software to enable pick-by-voice or kiosks to order food. It shortens the time spent completing tasks. Rhino Speech-to-Intent adds domain-specific voice control such as coffee makers. Leopard Speech-to-Text is used to build voice assistants with an open domain library, similar to Alexa. Porcupine Wake Wordor Cobra Voice Activity Detection is used to activate dormant devices and applications. Porcupine recognizes wake words, and Cobra detects voice activity to initiate listening for further commands. For example, to order food on a drive-thru kiosk or via phone Cobra and Rhino combination works perfectly well.

Menu Navigation

As products get complex, it gets harder to navigate within applications. UX designers work on alternatives such as hamburger menus, tab bars, and dropdown menus to help users. With voice-enabled menu navigation, users can get what they need by simply asking. The menu item they look for could be meeting notes on CRM, lab results in EHR or changing a password. Rhino Speech-to-Intent helps users find what they ask within a specific application and context. Porcupine Wake Word initiates voice control with a wake word. Commands like “Hey Pico, pay my credit card bill” can be powered by the Porcupine and Rhino combination.

Voicebots & IVRs

An IVR (interactive voice response) is a voicebot application that enables callers to interact with the host system through pre-recorded voice responses via a telephone keypad or speech recognition. Voicebots are used to offer a better service and improve agent productivity by triaging the calls. Rhino Speech-to-Intent is used to build IVRs on the platform of your choice to help users navigate without the help of an agent. Cobra Voice Activity Detection is used to initiate the conversation by detecting the voice activity in cases such as outreach calls. Cobra detects when the call is picked up and minimizes the idle time of agents.

More from Picovoice