AIsuru 系统集成商认证课程

AIsuru AI System Integrator Course Resources

Training material for the AIsuru System Integrator certification program by Memori.ai.

Versione italiana

Enrollment

Registration is not open yet. The link to the course catalog will be published here.

About this course

The course is aimed at developers, software engineers and technical staff who want to be certified as System Integrators for AIsuru, the conversational AI platform built by Memori.ai.

What you will learn

• Integrate AI agents into enterprise applications and workflows
• Connect AIsuru to existing business systems (CRM, ERP, databases, APIs)
• Implement authentication flows (SSO, Microsoft Auth, OAuth, JWT)
• Build integrations with MCP servers and advanced functions
• Deploy and configure AIsuru PaaS installations
• Apply security practices to production deployments

Who should attend

• Backend and frontend developers
• Software architects
• DevOps engineers
• Technical consultants
• IT staff working on AI integrations

Course modules

Format: 3 days of 6 hours each (online webinar), plus 1 day of live workshop.

Repository structure

system-integrator-course-resources/
├── SYS-03-integrating-enterprise-authentication/
│   └── manage_login/            # Authentication demo app
│       ├── README.md            # Full guide (EN)
│       ├── README-IT.md         # Full guide (IT)
│       ├── img/                 # Screenshots used by the guide
│       └── demo/                # Rails app, demos 1-4
├── SYS-04-advanced-functions/
│   └── manage_functions/        # Advanced functions demo app
│       ├── README.md            # Full guide (EN)
│       ├── README-IT.md         # Full guide (IT)
│       └── demo/                # Rails app, demos 1-2
├── SYS-05-web-component-programming/
│   ├── README.md                # Module overview (EN)
│   ├── README-IT.md             # Module overview (IT)
│   └── web_component/           # Web component / Dialog API demo app
│       ├── README.md            # Full guide (EN)
│       ├── README-IT.md         # Full guide (IT)
│       └── demo/                # Rails app, demo 1
├── SYS-06-mcp-server-integration/
│   └── manage-server-mcp/       # MCP server integration demo app
│       ├── README.md            # Full guide (EN)
│       ├── README-IT.md         # Full guide (IT)
│       ├── agents/              # System prompts and assets for demos 5, 7, 8
│       ├── img/                 # Screenshots used by the guide
│       └── demo/                # Rails app, demos 1-8
├── index.html                   # Demo hub opened by the launchers
├── start-all.js, stop-all.js    # Scripts behind npm run start:all and stop:all
├── avvia.bat, ferma.bat         # Double-click launchers for Windows
├── avvia.command, ferma.command # Double-click launchers for macOS
├── README.md                    # This file
└── README-IT.md                 # Italian version

Prerequisites

• Basic web development knowledge (HTML, CSS, JavaScript)
• Familiarity with REST APIs and HTTP
• Understanding of authentication concepts (OAuth 2.0, JWT, SSO)
• Docker Desktop installed, which includes Docker Engine and Docker Compose and is required by every demo
• Node.js installed, needed by npm run start in each demo and by the npm run start:all launcher in the repository root
• A code editor (VS Code recommended)

Getting started

Quick start with one double-click

If you would rather not use a terminal, download the repository and double-click the file that matches your computer, in the top-level folder:

The launcher checks that Node.js and Docker Desktop are installed, starts Docker Desktop if it is not running, prepares every demo and opens the demo hub. The first run can take about ten minutes: leave the window open until it says "Tutto pronto" (all set). The demos keep running after the window is closed. These launchers speak Italian, the language the course is delivered in.

The first time, your operating system may ask for confirmation:

• On Windows, if a security warning appears, click More info, then Run anyway.
• On macOS, if you see "cannot be opened because it is from an unidentified developer", right-click the file, then Open, then Open again. Once is enough.

If something goes wrong, the same folder contains avvio.log with the technical details to attach to a bug report.

Terminal setup

1. Clone this repository:

   复制

   git clone https://github.com/memori-ai/system-integrator-course-resources.git
   cd system-integrator-course-resources
   

2. Start a single demo from its demo/ folder:

   复制

   cd SYS-06-mcp-server-integration/manage-server-mcp/demo
   npm run start
   

3. Or start all of them at once, each on its own port:

   复制

   npm run start:all
   

   This builds and starts SYS-03, SYS-04, SYS-05 and SYS-06 in the background, then opens the demo hub (index.html) in your default browser. It behaves the same way on macOS, Linux and Windows, and requires Node.js in addition to Docker.

4. Stop everything with:

   复制

   npm run stop:all
   

Each module has its own guide with the platform-side configuration (agent credentials, MCP servers, tunnels). Read it before running the demo.

Demo ports

The ports differ from the Rails default (3000) so that all demos can run at the same time.

Module details

SYS-01: System Integration Principles

• Services to integrate and inventory of data sources
• How to provide context to the AI, and what the internals imply for the integration strategy
• Data accessibility and update frequency
• Limitations of RAG (Retrieval Augmented Generation) and vector databases

SYS-02: Frontend & Backend Integration

• Web component overview and embedding techniques
• API capabilities and integration patterns
• Enterprise use cases

SYS-03: Enterprise Authentication

How to connect enterprise authentication systems to AIsuru agents. The module ships a Rails demo app with four authentication approaches:

Learning objectives:

• Add authentication to an embedded AIsuru web component
• Use the LoginWithJWT API for programmatic authentication
• Configure a Trusted Application for backend-to-backend calls
• Configure Microsoft Azure AD / Entra ID SSO
• Build redirect-based authentication flows

Prerequisites: OAuth 2.0, JWT and SSO basics; Docker installed.

Module resources: SYS-03

SYS-04: Advanced Functions

How to write advanced functions that connect AIsuru agents to external REST APIs. The module ships a Rails demo app with two approaches:

Learning objectives:

• Create advanced functions in AIsuru and configure their webhook URL
• Configure webhook integrations with external REST APIs
• Import OpenAPI/Swagger files to generate functions
• Link functions to an agent and configure their usage
• Debug and verify function calls inside a conversation
• Handle function parameters and responses
• Apply function design and error handling practices

Prerequisites: REST API and HTTP basics; familiarity with the AIsuru web component (SYS-03); access to the AIsuru platform and agent credentials; JSON basics; Docker installed.

Module resources: SYS-04

SYS-05: Web Component Programming

How to program the AIsuru web component beyond plain embedding: let the agent answers trigger real actions on the page, and pass the page context to the agent. The module ships a Rails demo app.

Learning objectives:

• Embed and configure the web component with the credentials of an agent
• Trigger DOM actions from the agent answers using structured output tags
• Pass the page context at session start with initialQuestion
• Manage the dialog through the Dialog API and build a custom UI
• Case study: AI4furn

Prerequisites: HTML and JavaScript basics, familiarity with the AIsuru web component (SYS-03); Docker installed.

Module resources: SYS-05

SYS-06: MCP Server Integration

How to connect AIsuru agents to external data sources and services through the Model Context Protocol (MCP). The module ships a Rails demo app with eight demos, from the MCP servers built into the platform to custom ones:

Demos 4, 5, 7 and 8 run entirely inside AIsuru and need no local service. Demos 1, 2, 3 and 6 need Docker and an ngrok authtoken, which is pasted on the demo page: the Start demo button on each page starts the right tunnel, stops the others (a free ngrok account allows one active tunnel at a time) and shows the value to paste into AIsuru. The manual procedure is still available on the page as a fallback.

The ready-made system prompts and the downloadable assets for demos 5, 7 and 8 live in agents/ and are also served by the demo pages, since attendees do not have this repository.

Learning objectives:

• Understand the MCP architecture and the role of tools, resources and prompts
• Use the MCP servers built into AIsuru (MongoDB, MySQL, Scheduler, Persistence, Skills)
• Build and deploy a custom MCP server in Node.js
• Expose local services publicly with ngrok tunnels
• Expose an agent as an MCP server, and tell the Consumer token from the Builder token
• Turn an existing REST API into agent tools with the OAuth/API connector
• Combine several MCP servers on the same agent

Prerequisites: Docker and Docker Compose basics; familiarity with the AIsuru web component (SYS-03); an ngrok account for demos 1, 2, 3 and 6; Node.js basics.

Module resources: SYS-06

SYS-07: Automation Integration

• Meeting synthesizer implementation
• SharePoint and Microsoft 365 integration
• Enterprise workflow automation scenarios

SYS-08: Pre-Release Best Practices

• Red Teaming strategies for AI systems
• Domain whitelisting and CORS configuration
• Secure configuration from the AIsuru dashboard
• Authenticated functions (Bearer tokens, JWT validation)
• Data isolation policies per customer and context
• Audit logging and traceability of tool usage
• GDPR compliance and data retention policies

Documentation and resources

The Engine API covers session management, dialog, memories, intents, functions, webhooks and NLP. The Backend API covers user management, memori CRUD, integrations, import/export and tenant management.

About Memori.ai

Memori is an Italian AI company founded in 2017 and based in Bologna. It works on conversational AI with a focus on EU AI Act compliance, enterprise security and data protection by design. AIsuru is its platform for building, training and deploying conversational agents integrated with enterprise systems.

License

This material is provided for AIsuru System Integrator certification training.

© 2025 Memori S.R.L. All rights reserved.

Need help? See the documentation, the AIsuru platform or memori.ai.