Use BOTs to generate efficiency and a better experience on top of the Voice Channel

Risorsa 1

Click here to subscribe and partecipate

Eudata Webinar: VoiceBOT

Use Artificial Intelligence and speech recognition to drive automation on the Voice Channel

A human-like conversation between a BOT and a customer might seem like a futuristic promise, but in Eudata we have already turned it into reality thanks to the VoiceBOT module of Convy.AI.
The Voice channel is still the most used in interactions with customers, representing in fact about 70% of total traffic volumes. Focusing on this channel allows to address great efficiency, gaining also in the quality of the service provided.

Convy AI manages the voice channel with the same degree of intelligence and with the same decision-making capacity used for textual channels, allowing the creation of conversations in real time, bidirectional and contextualised, with the ambition to solve both simple and complex processes.
The VoiceBOT module exploits Convy’s centralized intelligence and makes it available for the Voice channel as well as Chat, Social, Mail and SMS.

Thanks to the integration of Convy AI with existing IVR or telephony systems it’s possible to:

  • Create an automatic experience through natural language for your customers
  • Reduce IVR navigation times, eliminating queue times with great economic advantage but above all with a significant value perceived by the user
  • In case of hand over to an operator, the conversation is optimized thanks to a better call classification

Why Convy.AI is the right choice:

The centralized decision-making intelligence of Convy.AI allows the creation of generic voice and text services. Convy in full autonomy adapts the conversational experience to the peculiarities of the various frontends.

Convy.AI can interact with different components enriching its information capital and being able to take the right decision
Convy.AI has a semantic engine that understands natural language and interprets its content according to context

Convy.AI is a complete conversational module, which guarantees front ends, backend integrations and administrative, monitoring and analytics modules

Convy supports the following IVR / Telephone Systems:





Recording ITA


Recording ENG


More suggestions for you



Success Case

New Release 4.7


Customer Centric Analytics


Proactive engagement





Lexical analysis

These 4 levels of analysis and comprehension of text are automated and integrated inside any solution based on the Cogito technology

Grammatical analysis

Syntactic analysis

Semantic analysis

Entities recognition

Automatic recognition of persons, products, companies, organizations, not based on closed lists but on inference


Recognition of language constructs as: homonymies, synonyms, hyperonymies, hyponymies…

Integration through the natural Language

By means of the functionalities described above, Cogito enables the platform to the interaction in the natural language, to give the user an advanced experience that is not based on pre-determined Patterns for dialogues recognition and  development

NLP Personalization

The interaction modality in NPL is customizable through Cogito Studio: this way the system is able to evolve through the changements of the customer needs

Supported Semantic Engine

Cogito Expert System (OEM)

Api.Ai NEW

Microsoft Luis

IBM Watson will be supported Soon

Supported languages

Supported languages

12 supported languages: Italian, English, French, German, Spanish, Arabic, Korean, Chinese, Hebrew, Polish, Dutch, Greek

Level of coverage of the Italian language

450.000 predefined concepts, 6 million relationships among concepts

Personalization of the semantic network

Through a specific tool (Cogito Studio)

Telco Language Package

The general vocabulary over the years has been extended with specific telecommunication market-oriented terminology and concepts.

Through Cogito Studio it is possible to train the system to recognize particular terms and concepts of a specific company and its products and services offer

Automatic categorization

Text classification on the basis of a standard taxonomy

Various taxonomies are available, based on the international standards of the classification of information: news, intelligence, crimes, cybercrime, finance…

Texts classification on the basis of a personalized taxonomy

Through Cogito Studio it is possible to train the software to classify texts on the basis of a personalized taxonomy (for instance, the tree of the triplets of a CRM system)

Possibility to personalize the classification rules

Cogito Studio provides a specific language (C-Language) suitable for customizing and controlling with precision the logics of classification of any text or document

Information Extraction

Extraction of concepts from a text

Cogito can extract and process from a text not only simple words, but accurate and free of ambiguity concepts (for instance “access point”, “data network”, “optical fiber”, “IP address”…)

Structured data extraction

The system can recognize, extract and normalize structured data (for example customer code, fiscal code, dates)

Entities extraction

The system can recognize, extract and normalize entities like persons, product names, companies, organizations, cities, geographic locations, addresses

Data normalization

The extracted information are automatically normalized and reorganized (for example, a name can be split in first name and last name). Potential custom normalizations are managed in a simple way by Cogito Studio

Possibility to customize the extraction rules

Cogito Studio provides a dedicated language (E-Language) to customize and control with precision the extraction logics of the entities. This way, as well as the standard entities supplied out of the box, Cogito allows the recognition of complex entities specific to the customer context or the context in which Cogito has to intervene

Sentiment analysis

Sentiment recognition

The system detects the sentiment used to write a text at the entire document level or single phrase level

Sentiment ranking

The sentiment is not classified through a binary mode (positive/negative) but on a scale of values

Sentiment on the Entities

In the future it is possible to extend the sentiment not on a document or phrase level, but on a Entity level. So, every extracted entity will be classified as associated to a sentiment or not

Chatbot and multichannel conversation management

Multichannel conversational Engine


Supporting Facebook Messenger


Supporting Twitter


Supporting Telegram


Supporting Skype

Both Skype free and Skype for Business

Supporting Amazon Echo


Persistent Chat management on web


Management of email conversations


Livechat management


Sms management


IVR/Voice Portal management


Chatbot customization

Design of chat flows

Through the administration tool it is possible to set more chat depending on the fruition channel

Design of the conversation flow

The design of the conversation flow is completely customizable through a graphic interface

Intent recognition

The recognition can occur in different modes: semantic analysis of the user messages, keyword recognition, pattern matching rules and regular expression

Multi-intent management

Available politics of multi-intent management through a pre-defined order or disambiguation asked to the user

Semantic hooks

Possibility to understand customer requests during the conversation and perform jumps of context in different parts of the conversational tree

Conditional blocks


Semantic blocks

Native functionalities available thanks to the integration with Cogito

Multimedia menu

Possibility to implement multimedia menù with with the use of images, videos, sounds

Multimedia node

Possibility to implement multimedia card to send customers information material

Extraction of structured information from the conversation based on semantic analysis

Native functionalities available thanks to the integration with Cogito

Extraction of structured information from the conversation based on keywords


Extraction of structured information from the conversation based on pattern matching and/or regexp


Categorization of conversations

Native functionalities available thanks to the integration with Cogito and its relative functionalities of categorization on one or more categorization tree, also personalized

Email dispatch


Web Service Connector

Possibility to interact with external systems and applications via Web Service SOAP or REST

Database Connector

Possibility to record on an external database the data and information collected by the Chatbot

Salesforce Connector

Possibility to interact with Salesforce Service Cloud for the customer recognition and handover

Skype for Business (S4B) Connector

Skype for Business Connector


Possibility to create and publish different version of the flows for development, testing and production

Multiple publishing

Possibility to publish the same flow on different channels

Authentication node NEW

Convy now supports Oauth2 authentication to let customer authenticate through the bot (Facebook only supported)

Multimedia node NEW

Multimedia Node added to configure outbound multimedia object such as links, video, images and complex objects (the combination of the above)

Carousel in menu NEW

It's now possible to configure carousel menu to add images to buttons

Handover management

Handover based on customer errors

Conversation forwarded after N attempts to interpret customer messages

Handover based on  service timetable

Possibility to offer a callback calling for chats occurred out of the service hours of the contact center

Handover based on agents’ availability

Possibility to offer a callback calling due to agents’ unavailability (prior integration with a system of dynamic check of the agents’ availability)


Survey management

Possibility to provide surveys directly by questions/answers handled in chat or through an external link of a survey management site


Administration console

Web Administration Tool for Chatbot management and administration

Conversation data logs

All the chat data are tracked and recorded: IP, customer, device, text of the conversation…

Real time analytics

It allows the visualization of the active chats and a quantitative measurement of the tool usage through the graphic representation “Live Statistics”


It allows the recording of all the conversations on a database and the distribution of aggregate data for statistics

KPI Performance

The system tracks the number of received dialogs, managed by the system and closed, managed by error, closed due to timeout, handover to agent. All the management times are tracked

Semantic and Navigation analytics

Possibility to visualize detailed analytics of the conversations including the navigation of the flow nodes and of the semantic recognition

Conversation navigation detail reporting NEW

It's now possible to retrieve the entire navigation history of each dialog with all the echanged text (Human2Bot and Bot2Human)

Configuration Import/Export NEW

It's now possible to export and import flows. This feature can be used in case of backups or in case of flows propagation to different companies

Admin UX review (4.7) NEW

Complete review of the Admin UI and UX

Qualitative reporting NEW

Convy adds new reporting tools to analyze the chatbot quality and to help administrators and supervisor to mantain it. Semantic analytics, confidence analytics, flow navigation detail and dialogs detail containing all the details of a single dialog

Company default configurations NEW

Default configurations per company added. Once configured new flows inherits standard configurations from this

Performance review NEW

Convy has been optimized in order to raise up the number of concurrent dialogs handled

Market recognition

Patent technology

The Cogito platform has an international registered patent

Analysts’ rating

Forrester Research in June 2016 has nominated Expert System as one of the first 10 companies all over the world in the solutions market of Big Data Text Analytics.

Gartner in 2015 included Expert System among the main companies in the solution sector of Enterprise Search.

Eudata received from Gartner a Honorable Mention in the Magic Quadrant 2016 for the CRM Customer Engagement solutions


©2015 Eudata | Via M.Gioia 55/C - 20124 Milano | P.IVA 12421000154
| TEL. +39 0245495098