M365 Agent

Prev Next

Overview

The Sight Machine M365 Integration empowers manufacturers to interact with their production data directly within the Microsoft 365 ecosystem. By leveraging AI-powered agents, users can perform natural language queries, retrieve key performance indicators (KPIs), and generate descriptive statistics without leaving Microsoft Teams, Excel, or PowerPoint. This integration bridges the gap between raw factory floor signals and the executive suite, enabling data-driven decision-making in real time.

Before you Start

Before attempting to use the integration, ensure the following requirements are met:

  • Site Configuration: The agent is currently optimized for the demo-bottling environment. Users should be aware of specific data quirks and limitations inherent to this demo site.

  • Permissions: Integration may require internal IT approval from your organization to allow the Sight Machine Agent to communicate with your M365 tenant.

  • File Readiness: For M365 Copilot features (Excel/PowerPoint), your document must be saved before the agent can initialize.

  • Initial Setup: Perform the "One-time setup" instructions below to install the agent in your Teams environment.

Feature Benefits

Integrating Sight Machine into your M365 workflow provides several key advantages:

  • Unified Workflow: Access production insights directly within the tools you use for meetings and reporting.

  • Mobile Accessibility: Use the Teams mobile app to check real-time machine status or basic production info from any location.

  • Instant Data Export: Effortlessly populate Excel spreadsheets with tabular data retrieved via natural language prompts.

  • Automated Analysis: Rapidly perform complex tasks such as identifying statistically significant differences in process parameters or running downtime regressions for morning meetings.

How to interact with the M365 Agent

This workflow shows how to interact with the M365 Agent in Teams and Excel Copilot—how to start a chat, mention the agent, ask questions, and use the structured table results (including adding them to an Excel sheet).

Chat with the agent in Teams

  1. Install the Sight Machine Agent in Microsoft Teams.

  2. In Teams, start a new chat with the Sight Machine Agent.

  3. In the Type a message text box, ask a simple starter question, for example: “What machines do you know about?”

  4. Send the message and give a moment while it scans data sources.

  5. In this example, it would reply with a long table listing machines that looks like the following:

Use the agent through Copilot in Teams

  1. Start a new Copilot chat session.

  2. In the Message Copilot text box, type “@Sight Machine Agent” to ensure you are chatting with the Sight Machine agent. You will know that you are chatting with the agent if the name appears at the bottom of the message box like this:

  3. Ask a question, for example: “What was the production volume for color 218?”

  4. If this is your first interaction with the Sight Machine Agent, it will first display a “Hi” message while it is working on the data collection for your question.

  5. It should then reply with something like the screen shot below:

Use the agent through Copilot in Excel

  1. Open Excel and launch M365 Copilot with a blank workbook.

  2. In the Message Copilot text box, start by entering “@Sight Machine Agent”

  3. Enter a question, for example: “How did production conditions today compare to yesterday?”

  4. Similar to Teams Copilot, if this is your first interaction with the agent, it will display a “Hi” message while processing your request.

  5. Once the analysis is complete, it will return a data table similar the following:

  6. To add the results to Excel sheet:

    • Click the + Add to new sheet link below the table, or

    • Copy and paste the table into your worksheet

Summary

The Sight Machine M365 & Teams integration transforms how manufacturing data is consumed and shared. By bringing AI-ready data into standard business applications, it streamlines the analysis process—reducing reporting cycles from weeks to minutes. Whether through real-time mobile updates or deep-dive regressions in PowerPoint, the integration ensures that accurate insights are always available where work happens. User feedback through thumbs up/down and text logs remains essential for the continuous tuning and improvement of the agent's capabilities.