Introduction

Discover & understand what consumers want with Intent Manager

The Intent Manager is the suite of tools that allows you to build an intent taxonomy, analyze the most popular consumer intents, optimize how well they are being fulfilled, and discover intents that can or should be automated. The Intent Manager is made up of the Intent Builder and the LivePerson Intent Analyzer.

The LivePerson Intent Analyzer automatically analyzes customer messages sent to human agents as well as automated bots, instantly identifying and recording customer intent through every turn in the conversation. It integrates seamlessly with the Intent Builder to provide a powerful unified system for intent modeling and management.

The Intent Builder makes it possible for brands to create their own intents and entities organized by domain.

Intent Manager starter packs are available to accelerate the development of Intent Taxonomy with both vertical specific and universal starter intents to enable our customers to quickly get started with an intent-driven business. Intent starter packs are available for cross verticals such as e-commerce and insurance for purchases and escalations.

Terminology

Term Definition
Intent A customer’s desire to change something to do with your brand from state A to state B
Domain Domain is a list of intents and entities, that can be matched to user training phrases (utterances) with an intent-recognition model
Pre-built domain A pre-built domain is a starting point for you to get going with Intents. The intent-recognition model is filled with predefined intents, that have been trained and activated for your convenience. The model has been developed using aggregated and anonymized data from several brands. While pre-built domains are functional, you will need to customize it for your specific needs to get the best performance.
NLU NLU stands for Natural Language processing. In this context NLU refers to the customized machine learning algorithms that are trained to understand the intent expressed in your conversations
Training phrases (utterances) Training phrases are example utterances for each intent in your custom taxonomy. Pre-built domains use generic training phrases to get you started. To improve your model you can replace these generic phrases with examples of real consumer messages from your conversation history.

Example Use Cases

Business Analyst

Challenge

An Airline doesn’t have a real-time way to understand customer contacts by intent.

How LivePerson's Intent Analyzer Helps

The analyst uses LivePerson's Intent Analyzer to discover, define, and enable the following high impact intents

  1. Customer doesn’t understand policy changes
  2. Customer can't find their discount code
  3. The discount code doesn’t work
  4. Customer wants to cancel their account

Outcome

The analyst enables detailed real-time insight into the problem by defining intents

Agent Manager

Challenge

There is influx in account cancellations related to the policy changes

How LivePerson's Intent Analyzer Helps

The Agent Manager uses the LivePerson Intent Analyzer dashboard to explore conversations with the intent “Cancel my account.”

He discovers that some agents are successfully able to redirect the conversation and prevent cancelation by offering a concession. He uses this insight to develop a new policy

Outcome

The Agent Manager is able to quickly roll out a new training which results in a lower account cancellation rate

Bot Builder

Challenge

The Bot Builder wants to use automation to help with influx of messages related to a policy change.

How LivePerson's Intent Analyzer Helps

The Bot Builder reviews the LivePerson Intent Analyzer dashboard and notices that the “can’t find discount code” intent is, both, a high volume intent and is easily automatable.

She creates a bot in Conversation Builder for handling these requests and creates an intent-based routing rule in the Conversation Orchestrator that automatically directs customers to the bot.

Outcome

The company is able to greatly reduce the volume of agent conversations related to the policy change.

Next steps

  • For more information on the Intent Analyzer and getting started quickly with intents, click here.
  • For a step by step guide of creating, evaluating and optimizing intents, click here.
  • For more information on the Intent Builder, click here