LoginSign Up

Chazt Documentation

Chazt DocsGuide

How Chazt Agent Works

When a client is onboarded to Chazt, the Chazt Agent does not automatically handle every customer conversation perfectly from day 1.

When a client is onboarded to Chazt, the Chazt Agent does not automatically handle every customer conversation perfectly from day 1.

Chazt Agent automation is a gradual setup and training process. The agent becomes more useful as the client adds proper knowledge base content, flows, templates, catalog details, Chazt Agent settings, intent mappings, and real conversation feedback.

Important expectation

Chazt Agent is not a magic switch where all customer conversations become automated immediately after onboarding.

At the beginning, the agent may require regular review, correction, and tuning. The business team should expect an initial training and improvement phase before the agent can confidently handle most customer conversations.

The quality of automation depends on the quality of the client setup.

How the Chazt Agent works

Every customer conversation is first received by the Chazt Agent.

For each conversation, the agent checks the customer message, understands the intent, reviews available business context, and decides the best next action.

The Chazt Agent may choose to:

  • Answer from the Knowledge Base
  • Trigger a Flow
  • Send a Template
  • Send a normal Message
  • Show Catalog or product information
  • Hand over the conversation to a human agent

The decision depends on the customer message, available business data, Chazt Agent settings, intent mapping, and confidence level.

What each business needs before expecting strong automation

Every client needs a proper setup before expecting strong automation results.

Knowledge Base

The Knowledge Base should cover all common customer scenarios.

Examples:

  • Product questions
  • Pricing questions
  • Delivery questions
  • Order status questions
  • Return and refund questions
  • Store timing
  • Store location
  • Payment methods
  • Warranty or service policy
  • Escalation cases
  • Frequently asked objections

If the Knowledge Base does not contain the answer, the Chazt Agent may not respond correctly or may move the conversation to human handover.

Flows

Flows should be created for structured customer journeys.

Examples:

  • New enquiry flow
  • Lead qualification flow
  • Order booking flow
  • Support request flow
  • Complaint registration flow
  • Appointment booking flow
  • Follow-up flow

Flows are useful when the customer interaction needs step-by-step handling.

Templates

Approved templates should be configured for repeated or business-critical messages.

Examples:

  • Welcome message
  • Payment reminder
  • Order confirmation
  • Delivery update
  • Follow-up message
  • Re-engagement message
  • Offer or campaign message

The Chazt Agent can send templates when the customer intent matches the use case.

Catalog

If the client sells products or services, the catalog should be properly configured.

The catalog should include correct product names, prices, descriptions, availability, and other useful details.

This helps the Chazt Agent answer product-related questions more accurately.

Chazt Agent settings

Chazt Agent settings should be reviewed for each business.

These settings control how the agent behaves, how confidently it should reply, when it should use the Knowledge Base, and when it should transfer the conversation to a human.

Incorrect settings may cause poor replies, unnecessary handovers, or missed automation opportunities.

Intent mapping

Intent mapping is one of the most important parts of Chazt Agent automation.

Each common customer intent should be mapped to the correct action.

Customer intentChazt Agent action
Asking priceReply from Knowledge Base or catalog
Asking product availabilityShow catalog or send message
Asking delivery detailsAnswer from Knowledge Base
Asking to place orderTrigger order flow
Asking for supportTrigger support flow
Asking for human agentHuman handover
Asking about offerSend template or message
Unknown questionHuman handover or fallback message

Proper intent mapping helps the Chazt Agent decide whether to answer from the Knowledge Base, trigger a flow, send a template, send a message, show catalog details, or hand over to a human.

Day 1 onboarding guidance

From day 1, the business should understand that automation will improve step by step.

The recommended onboarding approach is:

Day 1: Basic setup

  • Connect required channels
  • Add basic business information
  • Add common FAQs
  • Configure basic Chazt Agent settings
  • Set human handover rules
  • Add important templates
  • Create basic flows

How Chazt automation improves over time

Chazt Agent automation improves as real conversations are reviewed and the setup becomes stronger.

During the first few days or weeks, the business team should monitor conversations regularly.

They should identify:

  • Questions the Chazt Agent could not answer
  • Wrong or incomplete answers
  • Missing Knowledge Base content
  • Missing flows
  • Missing templates
  • Wrong intent mapping
  • Cases that required human handover

These findings should be used to improve the setup.

Continuous improvement

The Chazt Agent should be improved continuously based on real customer conversations.

The team should keep updating:

  • Knowledge Base
  • Flows
  • Templates
  • Catalog
  • Intent mapping
  • Chazt Agent settings
  • Fallback messages
  • Handover rules

Over time, the Chazt Agent becomes more accurate and can handle more conversations automatically.

Why human review is needed initially

Every business has different customer questions, language style, products, policies, and support rules.

Because of this, the Chazt Agent needs client-specific setup and training.

Human review helps identify:

  • What customers commonly ask
  • Which answers are correct
  • Which cases need automation
  • Which cases need human support
  • Which flows are missing
  • Which templates should be used
  • Which Knowledge Base articles need improvement

Without this review process, automation quality will remain limited.

Recommended client communication

During onboarding, clients should be clearly informed:

Chazt Agent automation improves gradually. On day 1, the agent starts with the available Knowledge Base, flows, templates, catalog, and settings. It may not handle every customer conversation perfectly at the beginning. The agent needs continuous training, review, and improvement based on real conversations. As the setup becomes stronger, the automation rate and response quality will improve.

Success checklist

Before expecting high automation, make sure the following are completed:

  • Knowledge Base covers all major scenarios
  • Common intents are mapped properly
  • Required flows are created
  • Required templates are approved and configured
  • Catalog is updated
  • Chazt Agent settings are reviewed
  • Human handover rules are clear
  • Fallback messages are configured
  • Client team is reviewing missed or failed conversations
  • Improvements are made regularly based on real chat history

Summary

Chazt Agent automation is a setup-driven and training-driven process.

Better input creates better automation.

For successful automation, every business needs:

  • Proper Knowledge Base
  • Proper Chazt Agent settings
  • Proper flow setup
  • Proper template setup
  • Proper catalog setup
  • Proper intent mapping
  • Regular review and improvement

Clients should not expect complete automation immediately after onboarding. The Chazt Agent becomes more effective as the client data, conversation history, and automation rules improve over time.