Get Started using AI for Service

Set up AI for Service and deploy your first virtual assistant.

Prerequisites

• Kore.ai account with AI for Service access
• Admin permissions for initial setup
• (Optional) Channel credentials for deployment

Platform Overview

AI for Service provides a unified platform for customer experience automation:

Quick Start: Virtual Assistant

Step 1: Create a Virtual Assistant

1. Navigate to Automation AI → Virtual Assistants
2. Click Create Virtual Assistant
3. Configure basics:
   • Name and description
   • Default language
   • Target use case (Customer Support, HR, IT, etc.)

Step 2: Define Intents

Intents represent what users want to accomplish:

Intent: Check Order Status
Description: User wants to know the status of their order
Training utterances:
  - "Where is my order?"
  - "Track my package"
  - "What's the status of order #12345?"
  - "When will my order arrive?"

Step 3: Build Dialog Flows

Create conversation flows for each intent:

[User: Check Order Status]
        │
        ▼
[Collect Order Number]
        │
        ▼
[Call Order API]
        │
        ▼
[Deliver Status Response]

Step 4: Add Entity Extraction

Extract key information from user messages:

Entity	Type	Examples
OrderNumber	Regex	#12345, ORD-98765
Email	Built-in	[e-mail]
Date	Built-in	tomorrow, next Monday

Step 5: Connect Integrations

Add backend integrations:

1. Go to Integrations → Actions
2. Select an integration (Salesforce, Zendesk, custom API)
3. Configure authentication
4. Map entities to API parameters

Step 6: Test and Deploy

1. Use the Test panel to simulate conversations
2. Iterate on training and flows
3. Deploy to channels (Web, Voice, Messaging)
4. Monitor performance in Analytics

Quick Start: Search AI

Step 1: Create Search AI Application

1. Navigate to Search AI
2. Click Create Application
3. Name your search application

Step 2: Connect Content Sources

Add data sources for search:

Source Type	Examples
Documents	PDF, Word, Excel
Knowledge bases	Confluence, SharePoint
Websites	Help center, product pages
Databases	Structured data

Step 3: Configure Retrieval

Set up search and answer generation:

Retrieval Settings:
  chunk_size: 500
  top_k: 5
  similarity_threshold: 0.7

Answer Generation:
  model: gpt-4
  citation_style: inline
  max_length: 300

Step 4: Test and Integrate

1. Test queries in the Workbench
2. Integrate via API or embed in virtual assistant

Quick Start: Contact Center

Step 1: Configure Contact Center

1. Navigate to Contact Center AI
2. Complete initial setup:
   • Business hours
   • Default routing rules
   • Agent capacity settings

Step 2: Set Up Queues

Create queues for different inquiry types:

Queue: Sales
Routing: Round-robin
Skills required: sales, product-knowledge
Hours: Mon-Fri 9am-6pm

Queue: Support
Routing: Skill-based
Skills required: technical-support
Hours: 24/7

Step 3: Add Agents

1. Go to User Management → Agents
2. Add agents with profiles:
   • Skills and proficiencies
   • Queue assignments
   • Capacity settings

Step 4: Configure Channels

Connect voice and digital channels:

Channel	Setup
Voice	Configure phone numbers, IVR
Web Chat	Embed widget, customize UI
WhatsApp	Connect WhatsApp Business API
Email	Configure email routing

Next Steps

Goal	Resources
Build complex dialogs	Flows & Dialogs
Deploy to voice	Channels
Connect CRM	Integrations
Set up agent assist	Agent AI
Enable quality management	Quality AI