Bot Development Overview
This guide explains how to shape your bot’s behavior in Irmaya, including tone, business context, and the conditions that should trigger human review.
Overview
Bot Development is where you define how the assistant should behave in customer conversations.
This area typically covers:
- AI behavior and tone
- business profile and operating context
- handoff rules for situations a person should handle
When to Use This
Use this section when you need to:
- set the assistant’s communication style
- define the business context behind responses
- decide when conversations should move to human handling
AI Behavior
AI behavior helps determine how the bot sounds and how it approaches replies.
Tone
Choose a tone that matches your brand and customer expectations.
Common directions include:
- professional
- friendly
- casual
Response Style
Decide whether responses should stay short or include more explanation.
This affects:
- how quickly customers get answers
- how detailed the assistant sounds
- how much context is included in replies
Language
Set the primary language the bot should use and review whether multilingual behavior is needed for your workspace.
Business Profile
The business profile tells the bot how to represent your company.
Useful profile elements include:
- business name
- short description
- operating hours
- service areas or specializations
The clearer this context is, the easier it is for the bot to stay aligned with your business.
Human Handoff Rules
Not every conversation should stay with the assistant.
Handoff rules help define when a person should step in, for example when:
- the customer explicitly asks for a human
- the issue is complex or sensitive
- the assistant is not confident enough to respond well
Configure Your Bot
When setting up the bot, work through the areas in this order:
- review AI behavior and tone
- define the business profile
- set initial handoff expectations
- test the configuration with simple customer scenarios
Keep the first version simple. It is usually better to refine gradually than to over-configure too early.
Best Practices
- start with a clear tone and a narrow use case
- keep business context accurate and current
- use handoff rules to protect quality, not to avoid human work entirely
- test changes after updating bot behavior
- use repeated failures as signals for what to improve next
Common Issues
Responses feel too generic
Improve:
- business context
- bot behavior settings
- knowledge quality
Too many conversations go to human review
Review:
- handoff thresholds
- confidence expectations
- missing knowledge or unclear behavior rules
Conversations never escalate
Check:
- whether handoff rules are configured
- whether the current thresholds are too loose
- whether sensitive cases are being identified properly
Related Guides
- Human Handoff Support — Review handoff workflow
- Knowledge Base Setup — Add business information
- Templates Overview — Reuse common setups
- Quick Start Guide — Complete initial setup
- Troubleshooting — Resolve common issues