Ir para o conteúdo

Overview of Jitterbit Harmony AI

Introduction

Jitterbit's Harmony platform is infused with AI capabilities focused on security, governance, and accountability. It provides AI assistants that enable no-code development, low-code integration of large language models (LLMs) and AI agent development, pre-built agentic AI, and the tools for you to design your own custom AI agents and intelligent applications.

Harmony AI assistants

Jitterbit's AI assistants are integrated into different components of the Harmony platform to provide contextual, natural-language-driven development and support.

LLM integration and AI agent development

Jitterbit provides a set of pre-built connectors for connecting to third-party LLMs and building custom AI-driven workflows. Jitterbit Integration Studio enables the design, deployment, and management of custom AI agents. You can use Integration Studio's graphical interface and available building blocks to orchestrate calls to a selected LLM, manage data transformations, and define the agent's workflow logic.

  • Amazon Bedrock: Registers tools or sends a prompt.
  • Azure OpenAI: Registers tools or creates a text completion or image based on a prompt.
  • Google Gemini: Registers tools or sends a prompt.
  • OpenAI: Registers tools or creates a text completion, transcription, image, or translation based on a prompt.
  • MCP Client: Lists tools or executes a specific tool on an MCP server.

Jitterbit Harmony is model-agnostic, giving you the flexibility to choose the best LLM for your needs. Certain connectors such as Amazon Bedrock provide access to additional models such as Anthropic, Amazon, Meta, and Cohere.

Jitterbit's many additional connectors facilitate direct API-based and standards-based connectivity, and include support for vector-based databases such as Pinecone and Elasticsearch. Such databases support smarter search to instantly retrieve the most relevant information from a large dataset.

Additional connectors such as Redis v2 connect to endpoints that are designed for managing short-term memory of natural and context-aware conversations, or you can use Jitterbit's cloud-native Cloud Datastore for both short-term memory and long-term persistence for continuity and personalization across sessions.

Pre-built AI agents (Marketplace)

Jitterbit Marketplace includes pre-built, configurable AI agents designed for specific business functions. These agents act as templates for common use cases.

  • HR Agent: Internal HR agent to manage workflows such as employee onboarding, training administration, IT requests, and software provisioning.
  • Knowledge Agent: Internal knowledge base retrieval agent that can interface with internal systems, documentation repositories, and storage locations.
  • Sales Agent: Internal sales-related data retrieval agent configured to query connected systems such as a CRM.
  • Reactive Agent: Basic AI agent for learning purposes that responds to user queries without memory or advanced tools.
  • Contextual Agent: Basic AI agent for learning purposes that builds upon the Reactive Agent by storing conversation history and maintaining context across multiple interactions.
  • Salesforce Q&A Agent: Basic AI agent for learning purposes that builds upon the Contextual Agent by adding a specific tool to address particular questions regarding Salesforce account details.
  • GitHub Agent with MCP: A basic AI agent that leverages the Model Context Protocol (MCP) to execute tools on an MCP server and act as a chatbot assistant.

Note

All except for the Reactive and Contextual agents use the Retrieval-Augmented Generation (RAG) technique, which combines LLM reasoning with access to external tools and data sources.

Customized AI agents and intelligent applications

To solve your specific core automation and orchestration needs, you can either build your own customized AI agents and intelligent apps, or we can build them for you:

  • Build your own custom AI agents and intelligent applications
    Build your own custom AI agents using Jitterbit Integration Studio and Jitterbit API Manager, and incorporate those AI agents into intelligent applications using Jitterbit App Builder. Refer to the documentation under How to build AI agents to get started.

  • Use Jitterbit's agentic AI services
    Jitterbit provides a professional services offering to build customized agentic AI for you. This service provides a technical team for the end-to-end design, development, testing, and deployment of custom AI agents based on specific organizational requirements. Connect with an AI expert via our Jitterbit AI web page.