Get fast, intelligent answers to your customers' queries automatically, driven by your own business data sources.
What is the AI Assistant Block?
The AI Assistant block allows you to create an automated loop where customers interact directly with your AI-powered support system.
The AI draws answers from:
FAQs
Custom Data Files
Fetched Links
Product Sync (for product-related queries on WhatsApp & Instagram)
It ensures customers get quick, relevant responses without needing human intervention — until necessary.
Steps to Create an AI Assistant Block in Journey Builder
Go to Journey Builder:
Navigate to the Journey Builder section from BIK Dashboard.
Choose BIK AI in the Right Panel:
On the right-hand side panel, select the BIK AI option.
Select AI Assistant Block:
From the list of available blocks, select the AI Assistant Block.
Configure the AI Assistant Block:
Click on Save:
After selecting the assistant, click on the Save button to store the configuration.
Connect Nodes:
Connect the Talk to Agent and Main Menu buttons to the respective blocks or flow nodes as per your use case. This will allow the assistant to route users to other interactions or agents based on their input.
Test & Publish:
Test the flow to ensure everything works as expected in testmode , and then publish your journey
How Does the AI Assistant Loop Work?
When a customer enters the AI Assistant block, they stay in a loop interacting with the AI.
The user will remain in the AI loop until an exit strategy is matched.
If no exit condition is met, the AI continues the interaction.
Exit Strategy for AI Assistant
The Exit Strategy ensures a smooth end to the AI conversation at the right time.
The AI looks for two types of triggers to initiate an exit:
Exit Commands:
When a customer:Expresses a desire to end the chat (e.g., says "Stop", "Talk to human", "End chat", etc.)
Requests human intervention.
Out-of-Context Queries:
If the customer asks something irrelevant to the AI’s trained data (out-of-scope questions).
Post Exit Handling:
If a non-keyword input is added after exit, the default flow is triggered or agent is assigned.
If a keyword input is added after exit, the corresponding flow continues.
What Happens After Exit Is Triggered?
Once an exit is triggered, the AI Assistant sends a button to the user.
The user must click the button to complete the exit.
Only after clicking the button, the AI will move to the respective next node in the journey.
If no button is clicked, the AI loop will continue.
Data Sources Used by AI
The AI Assistant fetches answers from multiple data points:
FAQs
Custom uploaded files
Fetched dynamic links
Products (for product-related queries on WhatsApp and Instagram)
Note: Data sync is automatic —example once you add an FAQ or upload new data, the AI will automatically use it.
Important Behaviours to Note
Unexpected Responses:
Sometimes, if no matching data is found, the AI might respond unexpectedly.No Image Analysis:
The AI cannot understand media — it only processes text-based queries.Quote Reply Support:
The AI now supports quote replies to text messages.Product Recommendations:
For product-related queries, products will be sent automatically on WhatsApp and Instagram using synced product catalogs.Custom Prompt Support:
You can add custom prompts to fine-tune AI behavior for special use cases.
How to Train the AI?
Add data sources
Data will automatically sync for AI use — no manual sync required!
Use of the AI Assistant
When a customer asks a question, the AI Assistant searches for the best answer by checking:
FAQs
Custom Data Files
Fetched Links
Product Sync
(For product-related queries on WhatsApp and Instagram, the AI fetches and shares products directly)
The AI matches the customer’s query with the available information and responds instantly, ensuring quick and relevant support without needing manual replies.
In short:
Customer asks → AI fetches answer from your synced data sources → Customer gets an instant response!
NOTE:
Out of Stock and Hidden Products:
Products marked as out of stock or bik-hide will not be shared by the Assistant when customers ask for product recommendations.
Conversation Context Awareness:
The AI remembers the conversation history during the session.
It can use previous context to generate better answers.
Handling Conflicting Data from Multiple Sources:
Since the AI pulls answers from multiple data sources (FAQs, files, etc.), there may be conflicts if two different sources have different answers for the same question.
In such cases, the AI may pick an answer from any one of the available sources.
It does not prioritize one source over another automatically.
Keep your data clean, updated, and avoid duplicate/conflicting entries across different sources to improve the AI’s reliability.
FAQs
1. What happens if a user doesn’t click the exit button?
The user will remain in the AI Assistant loop, and the AI will continue responding based on available data sources.
2. How do I stop the AI Assistant from continuing the conversation?
Ensure the Exit Strategy is properly configured — exit commands or out-of-context queries will trigger the exit flow. The user must click the exit button sent by the AI.
3. Can the AI Assistant analyze images shared by customers?
No. The AI can only understand and respond to text-based queries. Images are not analyzed.
4. How does the AI fetch product data?
Products are fetched via the synced product catalog and sent automatically during product-related queries, both on WhatsApp and Instagram.
5. Do I need to manually sync new FAQs or data for AI to use it?
No. Data sync is now automatic. Once you update or add FAQs/custom files, the AI will start using them immediately.
6. Can I customize the AI responses for special scenarios?
Yes! You can create and configure custom prompts under the AI Training section to control specific AI behavior.
7. What if the AI gives a wrong or unexpected answer?
If the AI cannot find a relevant answer in the data, it may generate a fallback or less relevant response. Regularly update and improve your training data to reduce such cases.
For further assistance or to raise feature requests related to AI Assistant, please contact [email protected].