Skip to content

Added in all of the changes from partnercenter-docs on Tools and Capabilities #71

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
jon-neher wants to merge 2 commits into
base: master
Choose a base branch
from
Open

Added in all of the changes from partnercenter-docs on Tools and Capabilities #71

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
e698aa6
Added in all of the changes from partnercenter-docs on Tools and Capa…
jon-neher Nov 6, 2025
3ee4408
Update AI documentation files with enhanced content
jon-neher Nov 6, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 4 additions & 0 deletions docusaurus/docs/ai/ai-capabilities/creating-custom-capabilities.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -87,6 +87,10 @@ Consider the prompt as instructions for when you hire an employee. Be specific a
3. Fill in the required tool fields (see [Tool Configuration Reference](#tool-fields-explained)).
4. Add multiple tools if your capability requires several API calls.

:::tip Detailed Tool Building Guide
For comprehensive step-by-step instructions on finding API documentation, using cURL import, working with AI assistants, and testing tools, see [Building Custom Tools](../business-app/ai/ai-capabilities/tools-overview/building-custom-tools).
:::

### Step 5: Test and Refine

1. Click **Save** to store your capability configuration.
Expand Down
87 changes: 72 additions & 15 deletions docusaurus/docs/ai/ai-capabilities/index.md
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,9 @@
---
title: AI Capabilities Overview
sidebar_label: AI Capabilities
description: Learn how AI capabilities work as specialized skills that enable AI employees to handle specific business tasks and automate customer interactions.
tags: [ai-capabilities, capabilities, ai-employees, custom-capabilities, automation]
keywords: [AI capabilities, capabilities, custom capabilities, AI employees, automation, tools, integrations, lead capture, appointment booking]
---

import { AISparkleIcon } from '@site/src/components/Icons';
Expand All @@ -9,13 +12,15 @@ AI Capabilities are specialized skills that define what your AI Employees can do

## What are AI Capabilities?

A **capability** is a discrete skill you can enable or disable for an AI Employee. Each capability contains instructions that guide the AI on:
A **capability** is a discrete set of instructions you can enable or disable for an AI Employee. Simply put, a capability is a section of prompt, along with an optional tool to call an API.

Each capability contains instructions that guide the AI on:

- Which actions to take in specific situations
- What outcomes to prioritize
- How to respond to different types of requests

Capabilities work together to create AI Employees that can handle complex, multi-step interactions while maintaining consistency with your business processes.
Multiple capabilities can work together to create AI Employees that can handle complex, multi-step interactions while maintaining consistency with your business processes. When the AI employee handles a request, it combines all the Capabilities into the prompt the governs its behaviour.

### Types of AI Capabilities

Expand All @@ -29,20 +34,14 @@ Business App provides pre-configured capabilities for common business functions.
#### Custom Capabilities
For specialized workflows, you can create your own custom capabilities that:

- Connect to external systems via APIs
- Simply add new instructions for the AI employee to follow (no tools needed)
- Connect to external systems via APIs (using tools)
- Execute complex business logic
- Integrate with third-party tools and databases
- Automate unique processes specific to your industry

## Business Impact

Capabilities transform customer interactions from generic chatbot exchanges into personalized assistance that feels like talking to a knowledgeable team member. Instead of customers waiting for business hours or getting generic responses, they receive immediate help with real business tasks like checking inventory, booking appointments, or getting account information.

This automation allows your human team to focus on high-value activities while ensuring consistent service quality regardless of volume or time of day.

## How Capabilities Work with AI Employees
## How capabilities work with AI Employees

### Capability Assignment
### Capability assignment
Each AI Employee can have multiple capabilities enabled simultaneously. The AI intelligently decides which capabilities to use based on customer requests and conversation context.

### Contextual Activation
Expand Down Expand Up @@ -70,6 +69,64 @@ For custom capabilities, tools define the technical implementation:
- Integration workflows
- Response processing logic

- **Configure Built-in Capabilities**: [How to Configure AI Capabilities](./configuring-capabilities)
- **Create Custom Capabilities**: [Creating Custom Capabilities](./creating-custom-capabilities)
- **Advanced Configuration**: Learn about prompt engineering and capability optimization
:::tip Learn More About Tools
Tools enable AI Employees to interact with external systems and retrieve real-time data. To understand the fundamentals of how tools work with APIs, see the [Tools & Integrations Overview](../business-app/ai/ai-capabilities/tools-overview). For a step-by-step guide to building custom tools, see [Building Custom Tools](../business-app/ai/ai-capabilities/tools-overview/building-custom-tools).
:::

## Testing and Optimizing Capabilities

Effective capabilities require iterative testing and refinement. Use these strategies to ensure your capabilities perform as expected:

### Testing Capability performance

**1. Test Trigger Conditions**
- Verify the capability activates when it should
- Ensure it doesn't activate when it shouldn't
- Try various phrasings of requests to test consistency

**2. Verify Information Gathering**
- Confirm the AI asks for required information before using tools
- Test what happens when customers provide incomplete data
- Ensure the AI handles edge cases gracefully

**3. Review AI Explanations**
- Check the reasoning behind capability activation decisions
- Identify when the AI chose one capability over another
- Use explanations to refine trigger conditions and prompts

### Optimization techniques

**Shorter is Better**
- Keep capability prompts concise and focused
- Avoid redundant or contradictory instructions
- Remove unnecessary details that don't affect behavior

**Use Examples**
- Include specific examples of when to use the capability
- Show examples of required information formats
- Demonstrate desired response patterns

**Format for Clarity**
- Use markdown headers to organize sections
- Use bullet points for lists of conditions or steps
- Use bold to emphasize critical instructions
- Structure prompts so they're scannable for both humans and AI

**Ask AI for Help**
- Copy your capability prompt and ask an AI to suggest improvements
- Request examples of edge cases you should handle
- Get feedback on clarity and completeness

**Test and Iterate**
- Make one change at a time
- Test the specific scenario you're trying to improve
- Monitor real conversations to verify improvements
- Be ready to refine further based on results

:::tip Performance Monitoring
Use the Explanations feature in Conversations to understand:
- Which capabilities are being used most frequently
- Where capabilities might be conflicting with each other
- When the AI chooses not to use a capability and why
- How well your trigger conditions are working in practice
:::
18 changes: 18 additions & 0 deletions docusaurus/docs/ai/index.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -10,6 +10,24 @@ import { AISparkleIcon, SettingsIcon, CRMIcon } from '@site/src/components/Icons

The AI Workforce is a set of configurable AI Employees that help do the heavy lifting for your business by taking on repetitive tasks so humans can focus on other work. Each employee uses assigned **Capabilities** and knowledge from the **Knowledge Base** to perform their work accurately and consistently.

## Understanding the core concepts

Your AI employees are powered by three key components:

- **Knowledge Base** provides the factual content your AI can search and reference when answering questions — like business hours, pricing, FAQs, and policies.

- **Capabilities** tell your AI employee how to act and when to take specific actions in conversations — like when to capture leads, how to qualify customers, or what tone to use.

- **Tools** connect your AI employee to external systems via APIs, allowing them to perform actions like creating appointments, checking inventory, or updating your CRM.

These three components can be mixed and matched to create powerful AI employees that handle complex customer interactions and automate business processes.

:::tip Quick Reference
- **Knowledge Base** = Information to reference (reactive, fact-based)
- **Capabilities** = Instructions on how to behave (proactive, process-based)
- **Tools** = Ability to take action in external systems (integration-based)
:::

## Key areas

<div className="row">
Expand Down
217 changes: 174 additions & 43 deletions docusaurus/docs/ai/knowledge-base.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,64 +2,195 @@
title: Knowledge Base Overview
sidebar_label: Knowledge Base
sidebar_position: 3
description: Understand how the Knowledge Base works with AI employees and the types of knowledge sources you can add.
tags: [knowledge-base, ai-employees, knowledge-sources, conversations-ai]
keywords: [knowledge base, knowledge sources, business profile, ai employees, conversations ai]
---

The Knowledge Base is where you centralize all the information your AI Employees need to provide accurate, helpful responses to your customers. Think of it as the shared library that all your AI team members can reference when answering questions about your business.
# Knowledge Base Overview

## What is your Knowledge Base?
## What is the Knowledge Base?

Your Knowledge Base is a collection of information sources that your AI Employees can access during conversations. This includes:
The **Knowledge Base** is where AI employees access factual information about a business. It ensures that answers are accurate and grounded in real data.

- **Business Profile Information** - Your basic business details, hours, location, and services
- **Website Content** - Pages from your website that contain relevant customer information
- **Custom Documents** - Uploaded files like FAQs, price lists, policies, or procedures
- **Manual Knowledge Entries** - Custom text you've written specifically for your AI team
:::tip Key Principle
Do not add behavioral instructions to the Knowledge Base. Use **capabilities** for mandatory behavior instructions and use **knowledge base** for facts that can optionally be looked up when needed.
:::

## How AI Employees Use Knowledge
## Why is the Knowledge Base Important?

When a customer asks a question, your AI Employee:
1. Understands what the customer is asking
2. Searches through your Knowledge Base for relevant information
3. Provides an accurate answer based on your business information
4. Falls back to general assistance if no specific knowledge is found
AI employees rely on the Knowledge Base to answer questions consistently and accurately. Without it, responses may be incomplete or outdated.

## Benefits of a Well-Maintained Knowledge Base
## How the Knowledge Base works

- **Consistent Answers** - All AI Employees give the same accurate information
- **Reduced Training Time** - New AI Employees instantly have access to business knowledge
- **Better Customer Experience** - Customers get precise, up-to-date answers
- **Time Savings** - Less need for human intervention in routine inquiries
When a customer asks a question, your AI employee:

## Best Practices
1. **Generates a search term** based on the customer's message (e.g., "What time do you close on weekends?" becomes "hours weekend")
2. **Searches all assigned knowledge sources** to find the most relevant information
3. **Retrieves the top 10 most similar results** from your knowledge base
4. **Includes this information in its response** to provide accurate, grounded answers

### Keep Information Current
- Review and update your Knowledge Base regularly
- Remove outdated pricing, policies, or service information
- Add new products, services, or procedures as they become available
This happens automatically for every customer interaction, ensuring your AI always has access to the right information at the right time.

### Organize by Topic
- Group related information together
- Use clear, descriptive titles for documents
- Create separate entries for different aspects of your business
:::info Technical Note
Behind the scenes, the knowledge base uses **vector embeddings** and **semantic search** to find relevant information based on meaning, not just exact word matches. This means the AI can find answers even when customers phrase questions differently than your documentation.
:::

### Write for AI Understanding
- Use clear, direct language
- Include specific details and examples
- Avoid ambiguous terms or internal jargon
## Types of Knowledge Sources

### Test and Iterate
- Monitor how your AI Employees use the knowledge
- Identify gaps where customers aren't getting good answers
- Add missing information based on common questions
You can add or update knowledge in several ways:

## Getting Started
### Business Profile
- Available for all accounts in **Business App**.
- Automatically added to AI employees.
- Shares the following fields with AI:
- Business name
- Address and/or service areas
- Website
- Email
- Phone number
- Categories
- Hours
- Services offered
- Short description
- Long description
- Social media URLs
- Booking URL
- Updated by editing the Business Profile.

To build an effective Knowledge Base:
### Text
- Add FAQs or custom information manually.
- Best for short, business-specific details.

### Website
- Crawl the business website to retrieve content.
- Ideal for keeping product, service, or policy information up to date.
- Can be refreshed manually or on an automatic schedule.

1. **Start with the basics** - Add your business profile and key website pages
2. **Identify common questions** - What do customers ask most often?
3. **Create targeted content** - Write specific answers for frequent inquiries
4. **Test with your AI** - See how well your AI can answer questions with the current knowledge
5. **Expand gradually** - Add more detailed information as you identify gaps
### File Upload
- Upload documents (PDFs, manuals, policies).
- Allows the AI to reference structured content not published on the website.

**Supported file types:**
- **Documents:** PDF, PowerPoint, Word docs, text files, Markdown
- **Structured data:** CSV, Excel spreadsheets (especially useful for pricing, services, specifications)
- **Code/Data files:** JSON, HTML (great for structured data)
- **Images:** Common image formats (JPG, PNG, etc.)

A strong Knowledge Base is the foundation of effective AI Employee performance. The time you invest in creating comprehensive, accurate knowledge sources will pay dividends in better customer interactions and reduced manual support needs.
## Knowledge, Capabilities, and Tools: What's the Difference?

## Knowledge base vs. capabilities vs. tools

The knowledge base is one of three key components that make AI employees effective:

- **Knowledge Base** provides factual information the AI can reference (what you're reading about now)
- **Capabilities** define how the AI behaves and when to take actions
- **Tools** enable the AI to perform actions in external systems via APIs

:::tip Understanding the Difference
The knowledge base contains **information to reference** (reactive, fact-based). In contrast, capabilities contain **instructions on behavior** (proactive, process-based), and tools provide the **ability to take actions** in other systems (integration-based).

**What it's for:**
- Business hours, location, contact information
- Pricing, product specifications, service details
- FAQs, policies, procedures
- Any factual information customers might ask about
:::

### Capabilities
**Instructions that shape behavior**

Capabilities tell your AI employee **how to act** and **when** to take specific actions in conversations. Think of these as process instructions you'd give to a human employee.

**Example scenario: Booking appointments**
- Your capability says: "Whenever someone asks a question, ask if they'd like to book an appointment"
- This makes your AI proactive, not just reactive
- The AI now consistently offers to book appointments, even when not explicitly asked

**What it's for:**
- Conversation flows and processes (when to capture leads, how to qualify customers)
- Tone and communication style
- Conditional logic (if customer asks X, do Y)
- Business rules and priorities

:::tip Knowledge vs. Capabilities
- **Knowledge** = "Here's the booking link" (reactive, information-based)
- **Capability** = "After answering questions, always offer to book an appointment" (proactive, behavior-based)
:::

## Best practices for Knowledge Base optimization

### Start with What You Already Have
Begin with existing resources rather than creating new content from scratch:
- Price lists or service catalogs (CSV or spreadsheet format works great)
- Product data sheets (PDFs)
- Your existing website content
- Training manuals or employee handbooks
- FAQ documents you've already created

**Example:** A pricing spreadsheet with columns for Service, Description, and Price is better than a paragraph describing all your services.

### Less is Often More
While you can add lots of knowledge, quality matters more than quantity:
- **Be selective:** Only add information that's actually relevant to customer questions
- **Avoid redundancy:** If the AI already knows general information, focus on what's specific to your business
- **Remove outdated content:** Less accurate information is worse than no information

**Example:** If you have 18 pages of blog articles about general topics, they might create noise. Focus on specific, business-relevant content instead.

### Knowledge is for Reference, Not Instruction
Don't try to control AI behavior through the knowledge base:
- ❌ **Wrong:** Adding "Always ask for email before phone number" to knowledge base
- ✅ **Right:** Put that instruction in a capability instead
- ❌ **Wrong:** Knowledge base entry saying "If customer asks about pricing, do X, Y, Z"
- ✅ **Right:** Add pricing information to knowledge; put the "do X, Y, Z" process in a capability

### Keep knowledge relatively static
The knowledge base works best for information that doesn't change frequently:
- ✅ **Good:** Service descriptions, product specifications, business policies
- ✅ **Good:** Pricing (if it changes monthly or less frequently)
- ❌ **Poor:** Daily specials or promotions (update via capabilities instead)
- ❌ **Poor:** Live inventory (use a tool/API for real-time data)

### Use website auto-refresh for dynamic content
If you do have website content that changes regularly:
- Enable the "Automatic refresh" option when adding website knowledge sources
- The system will check monthly for updates automatically
- This is great for blogs, menus, or service pages that evolve over time

### Exclude irrelevant website sections
When retrieving website content, use path exclusions to skip:
- Blog archives that are more for SEO than customer service
- Legal pages that aren't commonly asked about
- Admin or login areas
- Marketing pages not relevant to customer questions

**Example:** Exclude `/blog/articles/` if those posts are general content not specific to your business.

### Review what the AI actually uses
Use the conversation explanation feature to see:
- Which knowledge sources the AI referenced in responses
- Whether it found the right information
- If any knowledge sources are causing confusion

This helps you refine what to keep, update, or remove from your knowledge base.

## Frequently asked questions (FAQs)

<details>
<summary>How do I update the Business Profile knowledge source?</summary>

Edit the Business Profile in Business App. Changes automatically sync to AI employees.
</details>

<details>
<summary>When should I use text vs. website knowledge?</summary>

Use **text** for short, specific information not on the site. Use **website** for broader content that is already published online.
</details>

<details>
<summary>Does uploading a file automatically update if the file changes?</summary>

No. You'll need to upload a new version of the file to refresh its content.
</details>
Loading