Content Tagging

Content tagging is not enabled by default. See Enabling Tagging below to turn it on.

How to get there: Once enabled, go to Setup → Chatbot in the top menu → click your chatbot → Tags tab (under the Training section in the sidebar).

Overview

Content tagging allows you to organize your knowledge base content with categories (tags) and automatically show visitors only the most relevant information based on their needs. Instead of searching through all your content, your chatbot can focus on exactly what each visitor is looking for.

How It Works

1. Tag Management

Create Tags for your chatbot to categorize your content:

  • Tag Label: A human-readable name (e.g., "Belgium", "Netherlands", "Premium Support")
  • Tag Slug: An automatically generated identifier (e.g., "be", "nl", "premium-support")

Tags are unique per chatbot and can be used to organize both webpages and files.

2. Tagging Content

Assign tags to your knowledge base content:

  • Webpages: Tag individual crawled pages
  • Files: Tag uploaded PDF, Word, or text files
  • Text snippets: Tag training text entries

When you tag a webpage or file, all of its content chunks inherit those tags automatically. Content without any tags remains available to all visitors as general knowledge.

3. Pre-Chat Questions

Let visitors self-select their topic with interactive questions:

  • Create questions that visitors answer at the start of a conversation
  • Each answer maps to a tag
  • Display as buttons or a dropdown
  • Make questions required or optional
  • Add conditional logic — show follow-up questions based on previous answers

Conditional (Dependent) Questions

Questions can depend on a previous answer. For example, an insurance company could set up:

  1. Question 1: "Which insurance product do you need help with?" → Car, Home, Travel, Liability
  2. Question 2: "Which car insurance tier do you have?" → Basic, Limited Comprehensive, Fully Comprehensive
    • This question only appears if the visitor selected Car Insurance in question 1

This way, visitors only see relevant follow-up questions. Someone asking about travel insurance won't be asked about car insurance tiers.

4. Dynamic Filtering

Based on the selected tags, the chatbot searches only:

  • Content matching the active tags
  • General (untagged) content

Content with other tags is excluded from search results.

When Do You Need Tags?

Tags solve a specific problem: when your chatbot serves content that overlaps or conflicts between topics. Without tags, the AI mixes up answers from different contexts. If your content is naturally distinct (e.g., different products with unique names), the chatbot can usually tell them apart on its own — no tags needed.

Real-World Examples

Insurance Company: Multiple Products

The challenge: You offer multiple insurance products (Car, Home, Travel, Liability) with different terms, deductibles, and coverage details. When a customer asks "Am I covered?" - the answer depends on WHICH product they have.

Your setup:

  • Tags: car, home, travel, liability
  • Pre-chat question: "Which insurance product do you need help with?"

Customer experience:

  1. Customer asks: "Am I covered for windshield damage?"
  2. Pre-chat: "Which insurance product do you need help with?"
  3. Selects "Car Insurance"
  4. Chatbot answers based on car insurance policy terms specifically
  5. Doesn't confuse them with home or travel insurance details

Why this matters:

  • Each product has unique terms, coverage, and exclusions
  • Giving the wrong information could be costly or create liability issues
  • Customers get accurate answers for THEIR specific product

Event Company: Multiple Festivals

The challenge: You organize three music festivals with different lineups, schedules, bus routes, and ticket prices. A visitor asking "What's the bus schedule?" needs the answer for THEIR event.

Your setup:

  • Tags: summerbeats, electronight, harvest-fest
  • Pre-chat question: "Which event are you attending?"

Customer experience:

  1. Visitor selects "ElectroNight"
  2. Asks: "What's the bus schedule?"
  3. Gets ElectroNight shuttle info: late-night routes, return buses until 5 AM
  4. No SummerBeats or Harvest Fest info mixed in
  5. Asks "What's the refund policy?" — general untagged content comes through fine

International Service Platform

The challenge: You operate across multiple countries with the same types of content but different details per country.

Your setup:

  • Tags: netherlands, germany, france, australia
  • Pre-chat question: "Which country are you in?"
  • SDK integration: App can pre-fill country, only asks if unknown

Customer experience:

  1. User in Netherlands asks: "How do background checks work?"
  2. Selects "Netherlands"
  3. Gets answer with Dutch VOG (Certificate of Conduct) requirements
  4. Doesn't see Australian police check procedures

Multi-Location Business

The challenge: You have multiple store locations with different opening hours, addresses, and local services. A visitor asking "What are your opening hours?" gets a confusing mix of all locations.

Your setup:

  • Tags: store-amsterdam, store-rotterdam, store-utrecht
  • Pre-chat question: "Which store location?"

Customer experience:

  1. Visitor selects "Amsterdam"
  2. Asks "What are your opening hours?"
  3. Gets only the Amsterdam store info — no confusion with other locations

Tag Behavior

OR Logic for Multiple Tags

When multiple tags are active in a conversation, the chatbot uses OR logic:

Example: Active tags = ["be", "nl"]

  • Content tagged with "be"
  • Content tagged with "nl"
  • General content (no tags)
  • Content tagged with "de" is excluded

General Content is Always Included

Untagged content is treated as general knowledge and is always included in search results, regardless of which tags are active. This is useful for universal information like company info, general FAQs, refund policies, etc.

Tag Inheritance

When you tag a webpage or file:

  • All existing content chunks inherit the tags automatically
  • New chunks created from that webpage/file also get the tags

Best Practices

1. Plan Your Tag Strategy

Before creating tags, map out:

  • What content categories do you need?
  • How will tags be detected/selected?
  • What should remain general vs. tagged?

2. Keep Tags Simple

  • Use short, memorable slugs (be, nl, not belgium-country-specific)
  • Create only tags you'll actively use
  • Don't over-tag (tags should simplify, not complicate)

3. Balance Tagged vs. General Content

Tag only content that varies by context:

  • Tag: Country-specific regulations, pricing, product-specific details
  • General: Company mission, basic product info, universal FAQs, refund policies

Over-tagging can fragment your knowledge base. General content is available to all visitors.

4. Test Your Tags

After tagging content:

  1. Test conversations with different tag combinations
  2. Verify the chatbot finds the right content
  3. Check that general content still appears appropriately

Passing Tags Programmatically

Instead of using pre-chat questions, you can pass tags directly to the chatbot. This is useful when your application already knows the visitor's context (e.g., their country, language, or subscription tier).

Iframe Embed

Add tags as a comma-separated tags query parameter in the iframe URL:

<iframe
  src="https://app.boei.help/p/bot/YOUR_CHATBOT_ID?load_key=YOUR_LOAD_KEY&tags=au-pair,nl"
  width="100%" height="600">
</iframe>

Boei Widget — JavaScript SDK

Use the Boei() function to pass tags:

<script>
  Boei({tags: 'au-pair,nl'});
</script>

This sets the tags before the chatbot iframe loads. The tags are passed as a comma-separated string of tag slugs.

Boei Widget — Hidden Input

Alternatively, add a hidden input field to your page:

<input type="hidden" id="boei_prefill_customer_tags" value="au-pair,nl" />

Dynamic Example

You can set tags dynamically based on your application logic:

<script>
  // Set tags based on user's country
  var userCountry = getUserCountry(); // your logic
  Boei({tags: userCountry === 'NL' ? 'nl' : 'be'});
</script>

Tag Format

  • Tags must be comma-separated slugs (e.g., au-pair,nl)
  • Slugs are lowercase alphanumeric with hyphens only (e.g., premium-support, nl, au-pair)
  • Tags must match existing tag slugs configured in your chatbot

Enabling Tagging

Content tagging is not enabled by default. To turn it on:

  1. Go to Setup → Chatbot in the top menu → click your chatbot
  2. Open the Training tab in the sidebar
  3. Click Settings (gear icon)
  4. Enable "Enable content tagging"

If you don't have access to this feature on your current plan, you'll see a warning message. Contact support at [email protected] to get it enabled for your account.

Common Questions

Q: Can I tag content after it's been crawled/uploaded?
A: Yes! You can add tags to existing content anytime.

Q: Can a webpage/file have multiple tags?
A: Yes, you can assign multiple tags to the same content.

Q: What happens if I delete a tag?
A: Content tagged with it will become untagged (general content).

Q: Can visitors see tag names?
A: No, tags are internal. Visitors only see the pre-chat questions (if configured).

Q: How do I test if tagging is working?
A: Open the chatbot, select a tag via the pre-chat question, and ask questions that should only match tagged content. Verify the answers are scoped correctly.


Note: This feature requires activation. Go to your chatbot's Training → Settings to enable it, or contact support if it's not available on your plan.