In the era of Generative AI, having hundreds of thousands of learning assets is only useful if your users can actually find them. The Filtered Connective Intelligence API allows you to integrate Filtered’s connective intelligence layer directly into your own ecosystem, whether that’s a custom-built AI chatbot, Microsoft Teams, or a specialized internal portal.
By connecting to the API, you move beyond simple keyword searches and allow your tools to "understand" content relevance at a machine scale with human accuracy.
How the search available via our API works
The Filtered API doesn't just look for matching words; it uses a high-dimensional meaning space to find content that truly fits the user's intent. Read our Success Centre page on how the Filtered algorithm works to find out more (to be finished)
Key Features of the API Integration
1. Requests for Skills, Assets, and Pathways
Instead of a basic search bar that looks for keywords, the API acts as a "request engine." You can ask it to retrieve specific building blocks for your learning platform:
- Skills & Labels: You can request a list of skills relevant to a specific job role or business area. The API doesn't just return a name; it provides the full definition and "labels" (categories) so your system knows exactly what that skill means.
- Asset Retrieval: You can ask the API to find the best content (videos, articles, courses) for a specific skill. It evaluates hundreds of thousands of items in seconds to pick the ones that actually help someone learn that specific thing.
- Pathways & Curated Lists: You can request a "playlist" of content. The API looks at your requirements and automatically pulls together a sequence of assets that form a logical learning journey or "pathway".
How it works in simple terms:
- Relevance Scoring: Every time you request a skill or asset, the API gives it a "relevance score." This tells your system exactly how well a piece of content matches the skill you asked for.
- Threshold Management: You can set a "quality bar" (threshold). This ensures your platform only shows the highest-quality results and ignores anything that isn't a perfect match.
- Complex Requests: You can use "wordy" or natural language. If a user asks, "How do I become a Scrum Master?", the API translates that complex intent into a specific request for the right skills and the top 5 most relevant assets to get them started.
2. Skill & Profile Data Sync
The Filtered API can use role and skill data imported into Filtered from your HCM or created in Filtered. It can work alongside other APIs to do the following:
- Role Mapping: The connector knows the user’s job role and tailors search results accordingly.
- Learning History: It can be programmed to avoid suggesting courses the user has already completed.
3. Agent-to-Agent Connectivity
You can connect the Filtered API to other AI "agents" within your business. For example: Scheduling Assistants: the API can work alongside a scheduling agent to not only find a course but also book time in the user's Outlook calendar to complete it.
How do I get started?
Follow these steps to begin the implementation process:
- Credential Setup: To ensure a secure connection, Filtered needs to set your unique API credentials. Contact success@filtered.com to have these generated for your organization.
- Developer Documentation: Point your technical team to our official API Documentation. This contains all the endpoint details, authentication protocols, and search parameters required for integration.
Real-World Application: The "Learning Assistant"
ECITB has successfully integrated the Filtered API directly into their specialized learning environment to combat content chaos and provide a "Google-like" experience for the training providers that use their services. By utilizing the API, they have moved toward a "headless" architecture where Filtered acts as the intelligent engine driving search and recommendations behind the scenes.
Key API Implementation Features
1. Integrated Relevance Scoring
ECITB uses the API to pull metadata and project it directly into their front-end interface.
- Relevancy Bar: When a user searches for a topic (e.g., "Manual Handling"), the API returns a relevance score for each asset.
- Priority Sorting: The interface automatically sorts results by the highest relevance first, narrowing the user's focus and reducing the time spent searching.
2. Advanced Curation and Pathway Management
The API allows for complex operations that were previously manual tasks within the Content Intelligence platform.
- Automated Curation: The system can now add skills, change skill groups, and find relevant content programmatically.
- Pathway Generation: By using the API, ECITB can curate high-quality pathways in seconds, saving significant administrative time for the L&D team.
Real-World Case Study: The "Agentic" Approach
For this customer, a major UK financial services group, Filtered’s API serves as the intelligence layer for a sophisticated, chat-based Learning Assistant powered by Microsoft Copilot within Microsoft Teams.
By moving beyond a traditional platform interface, the customer has created an "agentic" workflow where the AI proactively manages the employee's development journey.
How the Learning Assistant Works
The assistant functions by connecting three distinct data layers to provide a seamless user experience:
- Context Gathering: The bot pulls real-time data from Workday, including the employee's specific job role and their learning transcript of completed content.
- Intelligence Layer: The bot calls the Filtered API to compare the user’s profile against the organization’s skills framework. It identifies precise gaps and requests the most relevant assets to fill them.
- Actionable Planning: Instead of just providing a link, the agent creates a personalized development plan directly within the chat interface.
Key Features of the Agentic Workflow
1. Automated Calendar Integration
One of the most powerful elements of this use case is the transition from "searching" to "scheduling".
- Time Blocking: Once a user agrees to a recommended course, the agent uses a Microsoft Outlook API to check the user's availability.
- Smart Booking: It automatically books a dedicated learning session in the user’s calendar based on the specific duration of the course provided by Filtered.
2. Multi-API Orchestration
The assistant doesn't rely on a single source; it uses the GenAI model to orchestrate multiple technical endpoints:
- Content Sourcing: It utilizes the Filtered API to source high-relevance assets.
- Profile Updates: The agent uses a separate Degreed API to ensure that once a plan is created or content is engaged with, the user’s external learning profile is updated accordingly.
3. Conversational Personalization
By leveraging the Filtered API’s ability to handle natural language, the bot can interpret "wordy" requests and provide contextually accurate responses based on the customer's internal standards.
The Outcome: From Platform to Partner
This "Agentic" approach demonstrates how Filtered moves L&D from a passive platform (where users must go to find content) to an active partner (where learning comes to the user).
- Reduced Friction: Users never leave Microsoft Teams to find, schedule, or record their learning.
- Verified Relevance: Every recommendation is backed by Filtered’s proprietary relevance technology, ensuring the bank's employees only see the highest-quality content for their specific role.
Read the Filtered API documentation
Ready to get started? Read the API documentation here