Contact us
To learn more about Personik opportunities for your team.
Conversational
UI and chatbots for business
It might sound surprising, but the first prototype of the modern chatbot was created in 1966! The Eliza program developed by Joseph Weienbaum was able to detect the key words in a user’s question and insert them into a template answer. More than 50 years later no one is surprised by the algorithms that allow you to control different devices using voice or text commands.
What is Conversational UI?

Conversational user interface helps the user to get all the required information using text or voice request. Earlier computers could recognize commands written in a programming language, and now artificial intelligence can also recognize natural language thanks to machine learning. The main concept of conversational UI is to make the communication between a user and a system easier.

There are two types of conversational UI: chatbots and voice assistants. Businesses use chatbots more frequently than voice assistants, because interaction by text messages is more familiar to users than voice commands.
How can one use chatbots?

Chatbots are mostly used in the B2C segment to reduce call center's response time by offloading simple requests to chatbots, leaving human operators for more complex requests. Bots use a knowledge base to create responses to requests, and users receive needed answers instantly.

It is also possible to organize support services for employees: they are also customers, but within the company.

The advantage of chatbots

Reducing costs
Investing in a bot allows you to optimize the costs of service centers: if standard issues are solved automatically, there is no need for a large number of line employees. Managers can focus on complex cases, and the level of service increases.
Ease of use
Chatbots can be integrated into messengers, applications and websites. Users are usually already familiar with those, so they don’t need to adapt to a completely new product.
Availability
Chatbots are available 24 hours a day. They save the user from a long wait on the support service line. A properly configured bot can answer a user’s question in a matter of seconds, which positively affects the user’s loyalty. Customers can contact the support service anywhere, when talking on the phone with a support employee is inconvenient.
Potential for improvement
Compared to applications created for the same task, a chatbot is more adjustable to companies changing needs. It is also easier to change the information required for answering users’ requests. It is done quickly, due to updates of the database and the introduction of new communication scenarios.
Data collection
A chatbot can collect the data about each user. That can be used to personalize the communication process. After analyzing the data, bots can provide users with suitable recommendations based on their previous requests. Higher accuracy of given answers improves user experience and increases the trust level.
What do you need to know if you want to implement chatbots into your business?

At first glance, it may seem that conversational UIs are simple enough and do not require deep elaboration, but they are not. Poorly constructed scripts, bad query recognition and lack of prompts only annoy users.

1
Focus on user experience
Creating a user-friendly chat bot is a non-trivial task. The bot should be able to satisfy the needs of its target audience. The sequence of steps in the bot script shouldn’t be long and complicated, and for the first-time users of a bot, a button interface is more convenient, because it allows them to navigate through available functions.
2
Artificial intelligence needs careful tuning
The main task of AI in bots is to transform text messages into a set of events, and it is possible only with a large knowledge base. Otherwise, AI will not be able to recognize requests, and users will get a negative experience from the bot.
3
Integration with informational systems
Users don’t like to wait for the operator’s response, so the chatbot should be as functional and convenient as possible. The key task of developers is to integrate the chatbots with various information systems to automate requests. Conceptually an interaction with a bot is structured like this: «A text request from the user — classification of the request — asking the user for specific details — calling a function from the informational system — responding to the user».

Who has already implemented chatbots?

Tinkoff Bank implemented the chatbot in 2017. Since then, bank customers have an ability to instantly receive the answer to any question related to the bank services. The company significantly reduced the cost of support, because the bot offloads the support operators by taking on most common requests. The tips that appear when clients start to type in a question help them to navigate through the bot's functionality.

Tinkoff chatbot receives 25-30 thousand requests on a daily basis, and ⅔ of the questions are answered automatically.
The chatbot of car sharing service YouDrive was launched in 2016, and today its clients can quickly receive the information about the service, rates and fines. They also can write to the support service in case of a problem: the bot independently distributes user requests to the responsible employees.

With this solution, YouDrive reduced their contact centre call volume by 2/3!
Chatbot of the pharmaceutical company WERTEKS was made for helping new employees in the adaptation process. The chatbot took on a lot of organizational work and relieved the newcomers, their mentors and managers from the routine. The functionality of the bot has expanded: with its help the company’s employees can get the answers to the frequently asked questions, find the essential contacts and receive corporate news.

Now 9 of 10 WERTEKS’ employees are satisfied with the tool.




Share the article in your favorite social network:

Let us know, how we can help!

Kate
Platform Architect
Edward
Software craftsman
Alex
Founder and CEO
Alexander
Master of Operations