IDfunction Agent
IDfunction Agent is EvolutionID's internal AI-powered platform that lets developers build, configure, and deploy PIAM systems through a conversational interface — instead of manually editing code and configuration files.
Instead of spending days on repetitive setup work for each new customer deployment, a developer opens a workspace, describes what needs to be done, and an AI coding agent reads, writes, and commits the right code — while the developer stays in control and reviews every change.
What it replaces
Without IDfunction Agent, setting up a new customer PIAM deployment involves:
- Manually forking a base template repository
- Hand-editing RBAC policies, badge workflows, and integration configs
- Writing database migrations and Docker configuration by hand
- Co-ordinating across multiple files with no automated validation
This typically takes 10–14 days per customer engagement.
What it provides
IDfunction Agent compresses that process into a guided conversation:
- Create a workspace — give it a name, pick an AI model, and connect it to a repository
- Describe the requirements — tell the agent what the customer needs in plain language
- Review and commit — the agent writes and modifies files; you review diffs and commit from the browser
- Deploy — push to GitHub and trigger the deployment pipeline, all without leaving the platform
Two systems
| System | What it is | Who uses it |
|---|---|---|
| IDfunction Agent (this platform) | The browser-based workspace where EvolutionID developers work with AI agents | EvolutionID developers |
| PIAM Product | The per-customer access control software the agent helps configure | Customer end-users |
This documentation covers the IDfunction Agent platform. The PIAM Product is the output — the code and configuration the agent helps produce.
Quick navigation
- Architecture Overview — how the platform is structured
- Creating a Workspace — get started with your first workspace
- Workspace Lifecycle — how workspaces behave when open, closed, or idle
- Using the AI Agent — how to work with the agent in a session