User Guide - Etendo Copilot - Getting Started

Overview

Etendo Copilot is a powerful tool integrated into the Etendo Classic interface, or accessible via API, that provides an efficient way to interact with assistants and use tools developed to solve specific problems. It is an innovative project designed to streamline your processes by harnessing the power of Artificial Intelligence. This page will guide you through the main features of Etendo Copilot.

Info

To install Etendo copilot, you can read the Copilot Installation guide in the developer's guide section.

Note

Remember that, to use this functionality, it is necessary to configure an API Key. For this, you can use one of your own, or you can contact the Etendo support team to purchase one.

What is Etendo Copilot?

At its core, Etendo Copilot is a groundbreaking initiative that redefines how developers and users interact with tools and information. It revolves around a central component, the Assistant which acts as the mastermind behind task delegation. This Agent has secondary modules referred to as Tools. The seamless communication between these components is facilitated via a RESTful API, ensuring a stateless and scalable interaction model.

Assistant

The Assistant serves you making on-the-fly decisions about which Tool is best suited to respond to a particular query. This intelligent decision-making ensures that you receive the most accurate and efficient assistance.

Each assistant has defined instructions, as well as the possibility to configure a knowledge base and a set of skills or tools. The assistant is able to make decisions based on a question, using the most appropriate knowledge base or tool set to answer a particular query.

In turn, it is possible to configure assistant managers, able to delegate a specific query to other specialized assistants and coordinate between the different assistants in your team to achieve an assertive response. This intelligent decision-making ensures that you receive the most accurate and efficient assistance. Assistants can be distributed in Etendo modules as datasets or created per environment.

Tools

Each tool represents a dedicated and self-contained project, meticulously designed to excel at specialized tasks. Whether it involves code translation, text analysis, or data manipulation, the diverse suite of tools operates in seamless coordination.

Info

For a overview of the available tools and their technical details, please refer to the Available Tools - Developer Guide.

Key Features

Effortless Integration: Etendo Copilot seamlessly integrates into your environment, adding an extra layer of intelligence to your workflow.
On-Demand Assistance: Send your queries to Etendo Copilot, and the Agent will guide you towards the most suitable Tool for the job.
Diverse Expertise: Our ever-growing selection of Tools covers a wide range of domains, ensuring you always have the right solution.
Open AI Assistants: Copilot is integrated with the Assistants technology developed by Open AI, allowing you to set up your assistants, trained with your own knowledge base, able to generate and interpret new code, and use the specific tools already distributed by Etendo or new ones.
LangGraph: In this case, this option works as a manager of other assistants and allows to select team members.
Langchain Agent: These assistants can perform specific tasks in natural language and provide contextualized responses, enabling the implementation of multiple AI models, the use of a proprietary vector database and internal memory management.
Multi-Model Assistant: This type of assistant can be used with multi-vendor models such as Anthropic and Gemini, in addition to existing OpenAI models.
Attach Files: Etendo Copilot allows users to attach one or multiple files for assistants.
Context Sharing: Etendo Copilot can automatically capture and share the context of your current Etendo Classic session, such as selected records or active windows. This eliminates the need to manually input information, streamlining workflows and improving efficiency.

Copilot Interface

Navegation Bar

In the Etendo Classic navigation bar, you'll find a Copilot icon that leads you to the chat pop-up.

Here, you can select an Assistant and engage in a conversation with it. Copilot facilitates communication with Langchain Agent, Multi-Model Assistant, LangGraph and Open AI Assistant types, for more information visit Assistant Windows documentation.

Note

By default, Copilot will open the last assistant previously used.

Tool Bar

In any standard Etendo Classic window, you will see in the toolbar a button with the Copilot icon. Clicking this button opens the Copilot chat, automatically passing the context of the current tab or selected records. This feature ensures that Copilot has all the relevant information about your current session, eliminating the need to manually input details and streamlining your workflow.

Context awareness

Etendo Copilot can take advantage of the context of the currently active window or record in Etendo Classic. Clicking the toolbar button updates the current context automatically.

Form Context: If you have a form open (for example, a sales order or a customer record), Copilot retrieves the key fields and any unsaved edits, offering you real-time suggestions or validations.
Grid context: When a row is selected in a grid, Copilot captures the details of those specific records, sending Copilot all relevant information so you can receive relevant AI-based recommendations.
Tab and window context: In multi-tabbed environments, context must be sent to Copilot each time you switch from one view to another (e.g., from Orders to Products).

Attach Files

Etendo Copilot allows users to attach one or multiple files for assistants to process. This feature supports various file formats, such as .pdf, .csv, .jpg, and more, enabling flexibility in use cases. However, the ability to interpret and process these files depends on the specific assistant configuration, the defined tools, and the underlying model used.

This functionality ensures that users can seamlessly incorporate external data into their workflows, whether it involves parsing documents, analyzing spreadsheets, or processing images. Assistants and tools can be tailored to address specific requirements based on the type of attached files, providing contextualized and intelligent responses.