Naumen Erudite

Naumen Erudite is a universal platform based on artificial intelligence to optimize client service.

Naumen Erudite provides comfortable human interaction with the robot. In addition to voice calls, Naumen Erudite can work with all popular text channels that take on a dominant role in the client service, including web chat, chat in a mobile application, instant messengers and social networks.

2023: Naumen Erudite VoiceOut Performance

On June 20, 2023, Naumen introduced a product - voice robots for automation of calls with per-minute charging Naumen Erudite VoiceOut. With the help of the solution, companies can quickly launch customer calls by bot according to individually developed scenarios and pay only seconds of conversation with subscribers. Automation allows large and medium-sized organizations to quadruple the cost of outgoing calls. As part of the product, Naumen experts configure and train turnkey voice robots, so they can provide conversion an average of 30% higher than contact center operators.

Companies solve a large number of problems in the field of outgoing communications with customers thanks to Naumen Erudite VoiceOut. Within two weeks, the solution can automate sales of goods and services, informing about promotions, changes in work or debts, confirmation and transfer of delivery, surveys loyalty and satisfaction, updating and bases more. The transactional model implemented in the product involves payment only in seconds of work, robot without rounding up or down.

Before launching the call, the Naumen Erudite VoiceOut team builds a dialogue script both for existing processes and tasks of the company and from scratch, and then continuously refines the bot, tests hypotheses to achieve the necessary KPIs and increase conversion. The integration of voice robots with internal information systems reduces the time before making a call. Automation of outgoing calls using the product shows the greatest efficiency in contact centers, where there are repeated tasks and regular campaigns for calling a base containing at least 50 thousand contacts.

Transferring the telemarketing feature to Naumen Erudite VoiceOut helps expand the sales funnel by reaching a larger base. The solution also allows you to improve your customer experience by making faster callbacks on incoming requests from customers. Online access to analytics in the product gives companies the ability to optimize operator staffing and marketing budgets.

The key feature of Naumen Erudite VoiceOut voice robots is performance even in fast start conditions. The experience gained in more than 40 projects in banks, retail and other industries made it possible to make voice robots, which show 90-130% of the operator's conversion. We have allocated work with outgoing calls in a separate direction in order to effectively solve specific call problems thanks to a combination of the capabilities of several Naumen products at once - Naumen Erudite, Naumen Speech AI and Naumen Contact Center. As a result, Naumen Erudite VoiceOut combines the speed of project launch with the depth of customer tasks, "said Sergei Popov, Director of the Naumen Contact Centers and Robotic Systems Department.

2021: Mosenergosbyt's bot assistant on the Naumen Erudite platform earned at Yandex.Alice

On June 17, 2021, Naumen reported that the bot assistant of Mosenergosbyt, created on the Naumen Erudite platform, learned to accept meter readings through Yandex.Alisa. The launched skill also allows customers to check the account balance, the existence of debt and the amount of the last payment. The virtual assistant of Mosenergosbyt talks about ways to pay for electricity and explains what to do in cases of lack of electricity or voltage drops.

You can activate the bot voice using the command "Alice, launch the Mosenergosbyt skill." The skill is available on all media and devices in which "Alice" is built mobile applications browser : in and "," Yandex as well as through the smart speaker "." Yandex.Station Customers will be able to communicate with AI the assistant (AI,) artificial intelligence voice and correspond in the chat. For a convenient and quick solution to the issue, the bot uses prompt buttons with answer options and links to Mosenergosbyt services in the dialog, for example, a personal account on the company's website.

The integration of the Naumen Erudite-based chatbot with the Dialogs "platform allowed Alice to be used as one of the call receiving channels and dialogue interface. Naumen Erudite receives questions asked through Alice, determines their meaning, asks the interlocutor for missing details within the script, and forms an answer that appears in the chat and is voiced by the voice assistant. The NAUMEN AI platform interacts with the internal information systems of Mosenergosbyt, which allows you to identify customers by personal account number and provide information about accounts and payments during the dialogue.

Mosenergosbyt keeps pace with the times and strengthens its position in the digitalization of remote customer service, "said Andrey Kovalev, General Director of Mosenergosbyt JSC.

Our partnership with Mosenergosbyt to automate customer service began in 2014, and together we strive to make the services in demand more accessible. Every month, voice bots based on Naumen Erudite receive more than 400 thousand calls from customers of the energy sales company, - said Andrei Zaytsev, director of the department of contact centers and robotic systems NAUMEN. - Launching a skill in a popular channel such as' Alice'gives customers the opportunity to quickly get advice or service in a channel convenient for them.

2020

Release of Incoming Call Robotic Classification Solution

On June 22, 2020, the company NAUMEN introduced a solution for robotic classifying incoming calls, which completely replaces - IVR menu contact center in and is installed on top of any communication platform. A product based on AI Naumen the Erudite platform speech technologies and a group of companies CST will help contact centers improve customer experience and reduce costs by reducing the classification time of each call by up to 20 seconds.

According to the developer, the solution will allow you to abandon multi-level IVR menus with tone navigation and proceed to preprocessing calls online, during which the robot clarifies the client's question and receives an answer in free form. By processing the request formulated by the client, the classifier robot determines the topic of the call and routes it to the target group of operators for further service.

"Our design experience shows that customers tend to perceive the IVR menu as an annoying obstacle to the operator, which forces them to spend time listening to all possible options, press the keys in tone mode and independently search for the type of their question among the proposed ones. This interaction format can negatively affect the level of customer satisfaction with the service, and our robotic solution is designed to make the call classification process convenient for customers, " noted' Andrei Zaytsev, Head of the Department of Contact Centers and Robotic Systems NAUMEN '

According to the developer, the NAUMEN solution speeds up the process of classifying cases and simplifies it for customers, contributing to an increase in the FCR (First Call Resolution) indicator. In addition, reducing the call preprocessing time to 20 seconds allows contact centers to reduce the cost of a multi-channel federal number 8-800 with per-minute charging. The dialog approach to classification, based on machine learning technologies, also reduces the number of errors in determining the topic of the call and incorrect routing of the call. Under these conditions, dedicated attendant skill groups will receive more target traffic, and the number of call transfers from one specialist to another is minimized. The accuracy of the classifier robot gives contact centers the opportunity to expand the number of allocated call topics, plan more detailed scenarios for their processing and create groups of operators with a narrower specialization.

As noted in NAUMEN, the classifier robot can work in conjunction with any popular telephony platform, and its deployment in a contact center takes from 3 to 6 weeks. This solution gives the business the opportunity to evaluate the technologies of dialog artificial intelligence in operation without significant resource costs and quickly get the first result. In the future, the classifier can become the foundation for a deeper robotization of service: the solution allows you to collect statistics on the popularity of client call topics and aggregate records of operators' dialogs with customers. Subsequently, statistical data will help select processes and scenarios, the transfer of which for processing to the robot will bring a good result, and the accumulated historical examples of service can be used to train a voice robot or chatbot.

"Especially for NAUMEN, we have developed a model of spontaneous speech in the telephone channel, which supports all the diversity of the Russian language, its dialects and accents. The experience and deep expertise of the CST group also allowed us to satisfy one of the current market demands: the safety of personal data and dialogues with clients. We combined all this in an on-prem solution that can be used in the customer's loop, which means that it provides good data protection, " noted' Dmitry Dyrmovsky, CEO of CST group'

The speech technologies of the CST provide the accuracy of recognition of spontaneous and integrated Russian speech with industry terms, and the Naumen Erudite NLU engine allows you to determine the client's intention with an accuracy of 85% and take into account the context of the entire replica of the interlocutor, which is necessary if there are similar service topics in the contact center. Companies will be able to manage the robotic classifier independently using interfaces that do not require programming skills and allow you to add new topics for classification, track online and offline performance, and further train the robot, according to NAUMEN.

Inclusion in the unified register of Russian software

On April 9, 2020, the company Naumen announced that Ministry of Digital Development, Communications and Mass Media Russia it had included Unified Register of Russian Software Naumen Erudite in its decision. The platform allows you to use technologies (artificial intelligence Artificial Intelligence, AI) to create dialogs robots and implement them into the client service of organizations.

Naumen Erudite contains a natural language engine and a set of tools necessary to implement all stages of the robotization project. Platform interfaces allow you to mark up data, train the robot to search and interpret intentions and facts in live speech, design service scenarios, track efficiency and control the quality of its work. At the same time, robot control is possible without programming skills, which increases the availability of AI technologies for the customer.

Thanks to to algorithms machine learning (Machine Learning, ML) and in-depth training (, Deep Learning DL), robots on the Naumen Erudite platform can freely navigate hundreds of topics, classify calls contact center to or support services, independently serve calls and chats with customers, and also achieve high conversion for outgoing calls. As of April 2020, NAUMEN robots are capable of operating in 11 popular remote service channels, including. social networks messengers

In the development of Naumen Erudite, we focus on one of the key trends in the use of intelligent technologies, extremely relevant for the public sector: the introduction of AI into typical business processes, thanks to which companies can switch living employees from routine tasks to more complex and optimize service costs. Clients also benefit from such transformations: some of the questions will help solve the robot, which will not have to wait long on the line or in the chat, and the operator will quickly connect to solving complex cases, noted Andrei Zaytsev, Head of the Department of Contact Centers and Robotic Systems NAUMEN

As of April 2020, robots on the Naumen Erudite platform determine the topic of appeals and track parcels at the Russian Post Customer Support Center, help residents of the capital find and return the evacuated car through the city contact center of DIT Moscow, advise clients of OTP Bank and Promsvyazbank, communicate with debtors in Mosenergosbyt and accept meter readings in many large energy sales companies.

The developer recalled that the register of domestic software as of April 9, 2020 includes 12 NAUMEN software products in the field of managing digital infrastructure, client communications and service.

2019: Features of Naumen Erudite and the challenges to be solved

According to the information for January 2019, Naumen Erudite allows you to solve a wide range of problems, ranging from the thematization of calls, to the complete automation of service provision:

• Classification of case topics followed by intelligent routing
• Operator Souffler
• Reception of readings of metering devices
• Ordering a taxi
• Payment of orders
• Virtual consultant on the website and in the mobile application
• Search for addresses of offices, branches, ATMs, etc.
• Check the current status of the order
• Robot collector
• Chat Surveys and Telephone Surveys

Features

• Advanced algorithmachine learning for free dialogue with a person in a natural language
• Special interface for training and configuring the robot, which does not require the participation of programmers
• Contact Center Automation Platform Integration Toolkit
• Ready-made dialog blocks for quick script setup
• The ability to learn from small amounts of data
• Omnichannel Mode Support
• Drag-and-drop Client Conversation Script Designer
• Ability to work from the cloud

Know-How and Pre-Configured Solutions

• Out-of-the-box language models built as part of completed projects for customers across industries
• Flexible adaptation of ready-made language models to unique customer terminology
• Creating new language models and working with various speech recognition engines without involving suppliers of ASR (Automatic Speech Recognition )/TTS (Text To Speech) systems
• Ability to work simultaneously with different types of data: voice transmitted (ASR) and DTMF (Dual Tone Multi Frequency) keys
• Understanding spoken language. For example, to the question "When to call you back"? the robot correctly perceives the answers "tomorrow after three," "by the end of the week," "the day after tomorrow," etc.
• The ability to extract facts from an open conversational replica. For example, from the phrase "Good afternoon, I applied for a tan the day before yesterday," the robot is able to extract the facts:
  • Document: Passport
  • When: the day before yesterday

• If the robot is not sure that it understood the question correctly, it can ask clarifying questions
• In case of discomfort, the client may ask to connect him to the operator at any time