Conversational Commerce has opened new channels for customers to interact with brands across all stages of their journey. With the growing number of customer interactions from different messaging apps, innovative companies are looking for ways to efficiently handle high-volumes of messaging conversations – while delivering a great customer experience.
LivePerson Conversational AI offers an end-to-end orchestration of brand to consumer conversations using Intent Analyzer for intent detection, Conversation Orchestrator’s AI tools for routing and recommending actions to agents, and Conversation Builder for building bots and automations. With LivePerson’s conversational AI capabilities, brands can personalize conversational experiences for each customer throughout their journey - at scale.
Using LivePerson’s powerful tools, Conversational AI has the ability to:
- Assess consumer intent and conversation context to predict what might happen next
- Route dynamically to human agents or bots to solve customer requests quickly
- Recommend the next best conversational actions, including asking clarifying questions, suggesting knowledge base answers, routing to a particular agent, and transferring to bots built in the Conversation Builder or in a third-party tool
- Give brands transparency into every AI decision so they can control and manage the way AI is used in their conversational pipelines - accelerating customer experience improvements and brand efficiency
How Conversational AI Works
Conversational AI offers an end-to-end orchestration of brand to consumer conversations using the following AI tools and automations:
Automatically identifies consumers intent in real time, through every turn in a conversation
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. The Conversation Orchestrator takes intent data and combines information from different data sources to identify the best intent match.
Intents can be built and trained in LivePerson, or uploaded from an existing system or other NLU engine. Brands can create intent dashboards to optimize operations, intent by intent, and build a data driven automation program.
Dynamic Routing by Conversation Orchestrator
Dynamic, personalized routing policies using intent, context and predictive attributes
Conversation Orchestrator is LivePerson’s AI engine that orchestrates every consumer conversation with a brand to an outcome they both value. Conversation Orchestrator works in the background to quickly understand the consumer’s intent, and route to human agents or bots to solve customer requests in the best way possible.
Conversation Orchestrator’s Dynamic Routing uses the consumer’s identified intent, and assesses the conversation’s context from various data sources, including: consumer attributes, operational metrics, and enterprise systems to route to the right agent.
By combining information from an external system like a CRM, the consumer can be immediately identified by name, location, and other information, such as VIP status. LivePerson Functions allows brands to easily integrate data from external systems, like Salesforce, using secure, LivePerson managed function services.
Using the Conversation Orchestrator’s workspace, brands receive a self-service experience to manage and operate the way that AI is used in their conversational pipelines ensuring they continue to deliver a great customer experience.
Recommends knowledge base articles and/or bots to the agent based on conversation context
Agent Assist analyzes all available knowledge base articles and bots for the identified consumer intent and recommends the best next actions, such as having a bot join the conversation in real time. Recommendations are made by choosing the highest ranked articles and bots by score; scores are based on machine-learned signals about the historical performance.
Agents can opt to manually use a knowledge base article to help the user or join a bot to the conversation, where they can monitor the conversation and remove the bot if needed.
- AI-based recommendations for bots that can fulfill customer intent at each turn of the conversation
- Join bots to conversations and use knowledge base articles
- Closed loop learning to improve recommendations based on agent feedback
Easily design, build and integrate bots and automation flows
Bots can be created in LivePerson’s Conversation Builder, or integrated as third-party bots by using a connector for Watson or Dialogflow. The Conversation Builder is fully integrated into LivePerson’s Conversational AI platform, allowing for seamless back and forth escalation with human agents and real-time monitoring by humans. In this way, LivePerson enables both human agents and bots to interchangeably handle consumer interactions to deliver the best consumer experience possible for each conversation.
Conversation Builder allows brands to build bots including everything from creating intents and building dialogues, to back-end integrations that connect to all consumer messaging channels and everyday systems like Salesforce.
LivePerson built the Conversation Builder as a dialogue-based tool for creating automations and bots at scale. This allows brands to leverage non-technical staff to build, optimize, and improve bots and automated conversation flows - creating bots that are more effective and competent, and drastically speeding up bot development and optimization.
Conversational AI Feature List
Conversation Builder Features
- Dialog management - no code
- Dialog management - code based
- Automatic context switching
- Slot filling
- Bot templates provide a way to quickly generate different dialogs
- Ability to inject code at every step of the interaction and also for processing user responses
- Channels - ABC, RCS, WhatsApp, Facebook, SMS, Web. bots support all channels supported by LivePerson
- IVR Support
- API Integrations for connection to your REST APIs, CRM, Billing, Inventory or other business systems
- Function as a Service (FaaS) support
- Authentication methods (e.g, OAuth 2.0, Basic Auth, mTLS, Tokens)
- Search based Knowledge Base
- Ability to use different NLU engines for training Knowledge Base
- Both English and non-English languages are supported. For non-English languages, external NLU engines are recommended
- CB native NLU - intent detection (text based)
- CB native NLU - entity detection
- CB native NLU - slot fill
- External NLU engine support for IBM Watson
- External NLU engine support for Google Dialogflow and Watson
- Developer user persona
- Non-technical user persona
- APIs for intent detection
- APIs for Knowledge Base query lookup
- APIs for bot analytics
- Measure bot performance using different reports:
- Messages, Conversations, Users, sessions
- Intent match, unmatch reports
- Custom event reports
- LivePerson NLU supports: English & Spanish
- IBW Watson supports: Arabic, Chinese, Dutch, English, French, German, Italian, Japanese, Korean, Portuguese, Russian, Spanish, and Swedish.
- Google Dialogflow supports: Cantonese, Chinese, Danish, Dutch, English, French, German, Hindi, Indonesian, Italian, Japanese, Korean, Norwegian, Polish, Portuguese, Russian, Spanish, Swedish, Thai, Turkish, and Ukranian. See all languages.
- LivePerson Private Cloud
- AWS Public Cloud
- Customer Private Cloud
Conversation Orchestrator Features
Conversation Orchestrator Workspace
- Self-service experience to configure Conversation Orchestrator features
- Analytics for Conversation Orchestrator features (e.g. availability, usage)
- Uses NLU to match intent and recommend a bot to join the conversation.
- Uses NLU to match an intent to recommend knowledge base articles.
- Integration of bots from LivePerson’s Conversation Builder, DialogFlow, and Watson.
- Create intent- and context-based (e.g., consumer profile data, wait time) routing policies.
- Apply actions - route/transfer to human agent, skill or bot.
- JSON-based, low code authoring
- API access to inbox system attributes (e.g., user, conversation & operational context).
- Create custom attributes with static data (e.g. list of emails)
- Create custom attributes with Functions as a Service variables (e.g, CRM integrations via FaaS)
- Carry over custom context through conversational journey
**Recommendation API **
- REST API to Conversation Orchestrator that exposes next action to concierge bot/LOB app for routing decisions
Conversation Orchestrator Concierge Bot
- Conversation Builder’s concierge bot template pre-built with intents and Recommendation API integration
Intent Analyzer Features
- Self-service intent building using Conversation Builder NLU
- Self-service intent building using third-party NLU service (i.e. DialogFlow, Watson)
- Self-service intent building using starter packs for a supported verticals (for brands with no existing intent modeling) - coming soon
Real-time intent detection
- Intent identification for human agent conversations (messaging only) & bots* (*with intents defined in Intent Builder)
- Intent Analyzer dashboard integrated with Conversation Builder
- Topline metrics & messages with intent
- Intent volume & confidence
- Intent time series
- Intent list with details
- Conversation quality by intent
- Conversation viewer with drill down
- Intent Analyzer professional services for intent modeling starter packs and advanced custom taxonomy
- Message level Intent Analyzer classification data available via public API