AI & Natural Language Processing
Integrate phone usage into your AI and NLP voicebots using SIP, WebSockets, or Dialogflow connector. Automate routine tasks and improve the user experience.
We Can Help When You
- Already have a voicebot and want to integrate it with phone usage
- Want to automate routine tasks
- Don’t have a voicebot but are eager to build it
- Want to upgrade your chatbot to a voicebot
- Want to improve your live agent performance
Built-in AI for simple scenarios
Standard voicebots incorporate AI in terms of speech synthesis and recognition only. This works well for cases such as appointment reminders, order confirmations, and simple feedback collection.
We allow you to choose AI from one of our built-in providers
Synthesizes speech in 27 languages, provides 197 total voice options, 100 of which are neural. Google’s technology is used in Google Assistant, Google Search, and Google Translate.
Provides 116 total voice options covering 35 languages and 49 dialects. These options include 36 neural options based on the latest deep learning technology.
Turns text into life-like speech. Polly's text-to-speech service covers 18 languages and 58 voice options including 13 neural ones.
Offers 41 voices with 19 neural options in 11 languages. IBM Watson solution can learn from customer conversations.
Uses deep neural network models for speech recognition and synthesis and is used to create a financial voice assistant called Oleg. It’s available in Russian.
Allows you to recognize or speak any text in 3 languages. SpeechKit is what powers Alice, the Yandex voice assistant.
Advanced NLP for intent recognition
Sometimes it’s not enough to just capture words that customers say. There are cases when you need to capture specific keywords from speech, recognize customer intent, and ask for missing information. This is where you need an advanced NLP voicebot.
We provide you with the Dialogflow connector
ES agent
is suitable for short conversations and has been praised for its simplicity. ES is often used with voice apps in which a short utterance matches one intent. For instance, a food delivery voice app where you can say: «I would like to order a pizza». The voicebot will offer available pizzas, ask for size, quantity and delivery address.
CX agent
is an advanced voicebot type that is suitable for complex and long conversations over 10 minutes. CX has two key features: voicebots can transfer calls to live agents and users can interrupt voicebots so that it starts to listen again.
Easily Integrate Your Voicebots
SIP
A classic interface that does not require installation of additional programs, although it lacks flexibility. SIP is limited by the RTP protocol so you need to ensure data you exchange with your bot is compatible with RTP. You can connect to IBM Watson voicebot using SIP.
WebSockets
A more flexible option that is compatible with all AI providers. WebSockets are advanced enough to support media streams, metadata, and control messages in the same channel.
Dialogflow connector
One of the leading solutions for voicebots is Google Dialogflow, and we have a one-click integration with it. Dialogflow is often considered the most powerful gRPC-based interface allowing you to recognize customer intents.
Avoid Voicemail and Connect Agents with Real People
If you perform outbound calling to provide notifications or reach out to prospective customers, you probably want to know whether a person or machine answers. If your call was picked up by a voicemail, it’s better to leave a pre-recorded message in the inbox.
When you make outbound calls, our answering machine detector checks if it’s a voicemail, answering machine, or live person on the other end. This allows you to spend time only on engaged individuals.
Our Prices
Speech recognition
per 15 seconds
Speech synthesis
per 10 characters
Speech recognition
per 15 seconds
Speech synthesis
per 10 characters
-
Default
-
Enhanced
-
Default
-
WaveNet
Speech recognition
per 15 seconds
Speech synthesis
per 10 characters
-
Default
-
-
Default
-
Neural
Speech recognition
per 15 seconds
Speech synthesis
per 10 characters
-
Default
-
-
Default
-
Neural
Speech recognition
per 15 seconds
Speech synthesis
per 10 characters
-
-
-
Default
-
Neural
Speech recognition
per 15 seconds
Speech synthesis
per 10 characters
-
Default
-
-
Default
-
Speech recognition
per 15 seconds
Speech synthesis
per 10 characters
-
Default
-
-
Default
-
Neural
Speech recognition
per 15 seconds
Speech synthesis
per 10 characters
-
-
-
Up to 1M characters
-
Over 1M characters
Audio input/output
Audio input/output
-
per 15 sec
-
-
-
Audio input/output
-
per 1 min
-
-
-
Key Features
Mobile, landline, virtual and toll-free phone numbers
Speech recognition/STT/ASR
Speech synthesis/TTS
Audio call recordings
Answering Machine Detection/AMD
Call transcriptions
Dialogflow connector