When even the knowledge base gets overloaded

When even the knowledge base gets overloaded

When even the knowledge base gets overloaded

When even the knowledge base gets overloaded

When even the knowledge base gets overloaded

How can customer service automation help?

How can customer service automation help?

How can customer service automation help?

Noémi Herczeg, VP of Mito Next, n.herczeg@mito.hu
July, 2020 · 5 min read

Status

  • Screen time has significantly increased
  • Customers use the interfaces in unusual ways
  • The demand for information increases at every touchpoint

Problem

  • Crisis / increased number of questions or questions on unusual topics, which are difficult to handle
  • We would like to / we need to get information to a large number of users, or even to the whole user base, at the same time

The world has changed a lot in 2020: the availability of products and services and the service channels used have all undergone a transformation. Usual activities might not be carried out the usual way anymore; the behaviour of customers has become different.

The world has changed a lot in 2020: the availability of products and services and the service channels used have all undergone a transformation. Usual activities might not be carried out the usual way anymore; the behaviour of customers has become different.

Managing the new situations brought on by these changes is of key importance.

Managing the new situations brought on by these changes is of key importance.

What are the first steps?

Managing expectations

If your customer service team is overloaded, admit it and direct traffic toward the self-service channels.

Updating the available knowledge base

‘Normal’ regulations need to be updated as soon as possible so that they respond to the current situation.

Reviewing the existing channels

We need to handle the situation there where our users go, on the channels they prefer.

How can FAQ automation help manage customer service overload?

The number of problems is significant but actually manageable.  Overloading often occurs because there is no time to deal with questions that would be otherwise easy to answer.

Attainable goal: by highlighting the 5 most frequently asked questions and using automation to answer them, the workload can be significantly decreased.

People prefer social channels more and more for reporting complaints

Usually they ask the same questions again and again

There is no capacity even for answering simple questions.

Possible solution:
Handling the 5 most frequent questions with a chatbot in customer service

Possible solution:
Handling the 5 most frequent questions with a chatbot in customer service

Possible solution:
Handling the 5 most frequent questions with a chatbot in customer service

Possible solution:
Handling the 5 most frequent questions with a chatbot in customer service

mol-chatbot-phones

Why?

Due to the noticeably increased traffic, comments will flood our social channels if we don’t handle them appropriately via chat. Nowadays there exist FAQ automation processes that can significantly help in directing the users to the appropriate channels in a cost-effective way, in case of certain keywords.

How?

How to select the most frequently asked questions?

  • Call Center logs
  • Chat(bot) logs
  • Web analytics
  • CRM, other databases
  • Transaction details
  • Social media interactions
  • Knowledge base contents

How to validate
user flows?

  • Expectation research
  • Qualitative research
  • Quantitative tests
  • Demo, prototype
  • A/B tests

How to choose a platform?

  • Reliance on CRM or business back-end system
  • Analytics needs
  • Need for omnichannel capabilities
  • The messaging platform that is most suitable from the users’ perspective

With the help of these tools, by answering these questions, you can build a working prototype of a customer support chatbot.

Less can be more!

It’s best to start with only a few functions and validate those. They should be introduced in the welcome message, demonstrating the bot’s functionality with simple examples. Help the users by showing them what they can enter and what they can expect an answer to.

Why can a chatbot in customer service be effective?

Better Decision Support

User tests can be carried out based on various use cases with the help of the prototype, without system integration, modelling actual operation.

Support for Multiple Channels at Once

A good solution can save the interactions from various channels into a single user profile.

Quick MVP

Quick MVP

You can create a “minimum viable product” that can go live and be tested in as fast as a few weeks, which saves you valuable time.

An MVP in weeks

But that’s just the tip of the iceberg...

Mitobot_illu2

Want to know more about
our chatbot solution?

Want to know more about our chatbot solution?

You can find everything in one place here:
mitobot.io

How to move forward?

You can read a detailed summary about Natural Language Processing here:
mito.agency/nlpinpractice