Development of AI widget for ArcGIS Experience Builder

Job ID: 39910572

Budget: $250 – $750 USD

I am considering creating a web map viewer that allows:

-Geospatial visualisation in ArcGIS Online (AGOL) with the locations of companies with investment potential.
-An AI assistant that answers questions about each company or asset.

According to the Esri community, there are already examples of GPT integration with Experience Builder, demonstrating that it is feasible to incorporate an assistant into a viewercommunity.esri.com. Therefore, we are looking for a developer with proven experience in Experience Builder Developer Edition to design a robust and documented widget.

Purpose of the contract
To develop a custom widget in ArcGIS Experience Builder (Developer Edition) that integrates an AI assistant. This widget must interact with the viewer data (company layers) and allow the user to ask questions in natural language. The assistant will respond with basic information about the company (sector, capitalisation, headquarters, indicators) and simple explanations for novice users, without giving investment recommendations.
Freelancer requirements
Demonstrable experience in developing widgets in ArcGIS Experience Builder (Developer Edition). Please attach projects or repositories.
Solid knowledge of React, ArcGIS API for JavaScript, and action/trigger communication patterns within Experience Builder.
Ability to integrate AI services (OpenAI or others) using HTTP requests and manage their response. Must configure API keys securely.
Handling geospatial data in AGOL: layers, fields, queries, and pop-ups.
Technical documentation skills in Spanish: you must provide a step-by-step tutorial explaining everything from installing Experience Builder Developer Edition to creating and deploying the widget, including configuring the AI wizard.


How to apply
Send your proposal including:
A brief description of your experience with Experience Builder and similar previous projects.
Links to repositories or demos.
Financial proposal and estimated time per stage.
Any additional suggestions you may have (e.g., handling sensitive data, ideas for optimising widget performance).