Service Overview
This training service is designed to help data engineers, analysts, and platform users build production-ready AI chatbots directly inside Palantir Foundry. Participants will learn how to design, develop, and deploy AI-powered chat interfaces that can answer questions, take actions, trigger workflows, and interact with Foundry datasets and ontology objects.
The session focuses on practical implementation, not theory. By the end of the training, participants will be able to create an AI chatbot that understands business context, queries Foundry data securely, performs actions, and integrates with Foundry applications such as Ontology, Contour, and Quiver.
This service is ideal for teams looking to build internal copilots, data assistants, or operational AI agents within their Foundry environment.
What You Will Learn in This Session
1. Foundations of AI Chatbots in Foundry
- Understanding AI chatbots vs AI agents in Foundry
- Where AI fits in the Foundry architecture
- Foundry components involved:
- Datasets
- Ontology
- Code Workbooks
- Contour Apps
- Security, permissions, and data access considerations
2. Designing a Chatbot Use Case
- Identifying high-value chatbot use cases:
- Data Q&A assistant
- Operational task assistant
- Alert and notification bot
- Defining chatbot scope, actions, and guardrails
- Designing prompt structure and system instructions
3. Integrating LLMs in Palantir Foundry
- Using LLMs within Foundry (Foundry-hosted or external)
- Prompt engineering best practices
- Managing conversation memory and context
- Handling structured vs unstructured responses
4. Connecting Chatbot to Foundry Data
- Querying Foundry datasets using SQL and PySpark
- Reading analytical tables and metrics
- Using ontology objects for semantic queries
- Converting natural language questions into data queries
5. Building AI Agent Actions
Participants will learn how to enable chatbots to perform actions, not just answer questions:
- Trigger pipeline runs
- Create or update ontology objects
- Send notifications (email, alerts, tasks)
- Execute business workflows
- Validate user intent before action execution
6. Creating a Chatbot UI in Contour
- Designing a chatbot interface using Contour
- Handling user inputs and responses
- Displaying tables, charts, and summaries in chat
- Error handling and fallback responses
8. Testing, Deployment & Monitoring
- Testing chatbot responses and actions
- Handling edge cases and failures
- Deploying chatbot to production
- Monitoring usage and performance
9. End-to-End Hands-On Project
During the session, participants will build a fully working AI chatbot, such as:
- A Sales or Operations Data Assistant
- A Project Status Chatbot
- A Data Quality Monitoring Bot
This chatbot will:
✔ Understand natural language
✔ Query Foundry data
✔ Interact with ontology objects
✔ Trigger actions and notifications
Who Should Attend
- Palantir Foundry Developers
- Data Engineers & Analytics Engineers
- Platform Engineers
- Solution Architects
- Business Analysts working in Foundry
Prerequisites
- Basic understanding of Palantir Foundry
- Familiarity with SQL or PySpark
- No prior AI or ML experience required
Outcome of the Session
By the end of this service, participants will:
- Confidently design AI chatbots in Foundry
- Build AI agents that perform real business actions
- Deploy secure, production-ready chatbot solutions
- Understand best practices for AI governance in Foundry