Introduction

This guide describes the end-to-end delivery workflow for M42 Assist – who does what, from initial sales scoping through to a fully configured customer deployment. For what the app does, see Solution description; for the architecture behind it, see Technical Documentation.

1. Process Architecture

The M42 Assist app in Microsoft Teams acts solely as a communication interface (gateway). The final implementation scope depends entirely on the customer’s business needs and follows a three-tier dependency model:

• Layer 1 (technical foundations – ALWAYS REQUIRED): An account on the CAI platform (new tenant provisioning) + Tenant ID authorization from Microsoft Teams.
• Layer 2 (ESM Notifications): Required if the customer wants to track Matrix42 Pro ESM tickets within Teams (importing XML templates and data card listeners).
• Layer 3 (AI Core): Required if the customer wants automated answers or AI agents powered by AI Core (Knowledge Discovery / Agentic AI + connecting knowledge sources and LLM configuration).

2. Step-by-Step Implementation Workflow

Step 1: Customer Scoping & Requirements Gathering

• Responsible Role: Sales / Pre-Sales
• Objective: Define the practical scope of M42 Assist for the customer to ensure accurate project estimates and smooth delivery.

Qualification Questionnaire (Sales Intake Form)

Before submitting an onboarding ticket to the Service Desk, Sales must obtain answers to the following 4 questions:

What is the primary objective of implementing M42 Assist?

• ITSM Notifications only (ticket tracking and status alerts)
• Intelligent AI Helpdesk / Conversational Knowledge Base (Knowledge Discovery)
• Autonomous actions and self-service workflows (Agentic AI)
• Hybrid (Notifications + AI capabilities)

Microsoft Teams Environment:

• What is the organization’s Tenant ID? (See Configuration, section 2, for how to retrieve it via the view_teams_info command).
• Has the customer’s Microsoft 365 Administrator allowed/approved the installation of M42 Assist in the Teams Admin Center?

Ticket Notifications (ESM Integration):

• Should the bot send private 1:1 notifications (to the ticket reporter / assigned agent)?
• Should notifications be routed to shared Microsoft Teams channels (Support Groups)?

AI Requirements (if AI capabilities are selected):

• Selected LLM Engine: Local Matrix42-hosted model or customer-managed cloud (e.g., Azure OpenAI)?
• Where is the organizational knowledge located? (Confluence, PDF manuals, ESM knowledge base articles, internal wikis)?

Step 2: CAI Tenant Provisioning (ALWAYS REQUIRED)

• Responsible Role: Professional Services (PS Team)
• Reference Documentation: Configuration, section 2 (CAI Platform Configuration); condensed checklist in the Internal Quick Start Guide, Steps 1–2.

Procedure:

• Log in to the master CAI Administration panel (PS Team responsibility - if there is no CAI platform installed, please send a ticket to Polish PS).
• Provision a new tenant for the customer:
  • Enter the customer name and create administrative credentials.
  • Create two dedicated Spaces:
    • Personal Space: Paste the customer’s Tenant ID collected during the intake phase (Note: this value must be globally unique across the database!).
    • Public / Channel Space: Select the public space created in step 1 to handle general channel communications.
• Verify baseline communication: Send a test message in Microsoft Teams to ensure the bot establishes connection.

Step 3: Notification Track — Matrix42 Pro ESM Integration

• Prerequisite: Customer confirmed the need for ticket notifications in Step 1.
• Responsible Role: PS Consultant
• Reference Documentation: Configuration, section 3 (ESM configuration); condensed checklist in the Internal Quick Start Guide, Step 3.

Step 4: AI Track — AI Core & Knowledge Discovery / Agents Configuration

• Prerequisite: Customer confirmed the need for Knowledge Discovery or Agentic AI.
• Responsible Role: AI Core Team / Delivery Consultant
• Reference Documentation: See the Knowledge Discovery and AI Agents sections of Solution description for the feature overview – no dedicated AI Core / Knowledge Discovery configuration guide exists yet in this documentation set.

Procedure:

• AI Core Environment Provisioning:
  • Deploy an AI Core instance on the customer’s environment (Submit a ticket to the Service Desk requesting AI Core enablement).
• Connecting the Language Model (LLM):
  • Based on the intake agreements: configure the connection to the locally hosted Matrix42 model or provide API credentials for the customer’s Azure OpenAI instance.
• Configuring Knowledge Sources (Knowledge Discovery / RAG):
  • Set up indexing connectors: ESM article databases, Confluence spaces, SharePoint document repositories, or PDF procedural guidelines.
  • Run the initial vectorization and content indexing pipeline.
• Agentic AI (if purchased):
  • Configure autonomous agent actions (e.g., self-service password resets, hardware status queries, external webhook triggers).
• Connecting AI Core to CAI:
  • In the CAI Panel, link the configured AI Core service to the customer’s project so that user inquiries in M42 Assist are automatically routed to the RAG engine.