How companies benefit from Microsoft 365 Copilot

With Microsoft 365 Copilot, Microsoft integrates artificial intelligence directly into the Microsoft 365 environment. But what can Copilot do, and what do companies need to consider? An overview of options and first steps.

With Microsoft 365 Copilot, generative artificial intelligence (GenAI) is moving into the familiar Microsoft environment for businesses of all sizes. The AI assistant integrates into Office applications such as Word, Excel, PowerPoint, Outlook and Teams. It is also available in the browser and has access to documents. In this way, Microsoft 365 Copilot simplifies numerous routine tasks in companies, promising an increase in efficiency.

There are several Copilot offerings for organisations:

• Microsoft 365 Copilot Chat is available free of charge and provides a chatbot in the browser.
• Microsoft 365 Copilot Business, which requires a licence, is integrated into Office applications and can access documents stored in SharePoint and OneDrive for Business.
• Copilot Studio is a low‑code development environment that enables organisations to build AI agents, automate workflows and develop topic‑specific chatbots.

Examples of how Microsoft 365 Copilot is used in companies

In the Work Trend Index, Microsoft asked early users of Microsoft 365 Copilot about how it impacted their day-to-day work. Seventy per cent said they had become more productive and up to 30 per cent faster in completing routine tasks. Meaning, it was a clear gain.

Microsoft 365 Copilot benefits employees who mainly do screen work and juggle documents, data and e-mails on a daily basis.

Copilot can support employees in their day-to-day work with such tasks:

Savings in time and efficiency

• Formulate e-mails and other texts based on keywords, rewrite them in a different tone and generate replies
• Compile a personal daily overview of appointments, e-mails and documents
• Create a PowerPoint presentation based on a document, such as a Word file
• Summarise e-mail communication with a specific person, including links to mentioned documents
• Summarise texts and presentations, even in another language
• Summarise transcribed (translated into text) online meetings and extract information such as schedules, tasks, etc.

Idea generation

• Create a structure and agenda for a workshop
• Generate ideas and brainstorm
• Coach for generating and refining ideas

However, you should always check the result of such tasks, because AI assistants do occasionally make up (‘hallucinate’) facts, and the result can be flawed.

Where Copilot does (not yet) deliver measurable value

As great as the potential of Microsoft 365 Copilot is, its use does not make sense in every scenario. The benefits depend heavily on the underlying conditions:

• Highly standardised processes: in clearly defined workflows, such as accounting or data entry forms, Copilot offers limited additional value.
• Unstructured or poorly maintained data: if documents are hard to find or spread across local and cloud storage locations, Copilot cannot realise its full potential.
• Unclear access rights: overly broad or unclear permissions reduce effectiveness and can even increase risk.
• Lack of training and user competence: without concrete use cases and basic AI know‑how, Copilot often remains unused. Training and change management are key.

Copilot and data protection

A key difference between Microsoft 365 Copilot and the solutions for private individuals is Microsoft’s access to the data. In the private solutions, Microsoft can use the information entered to train its language models. With Microsoft 365 Copilot, data remains within the organisation’s tenant and is not used to train large language models.

Because Microsoft 365 Copilot can access business documents in SharePoint, you should regulate the access rights for data security and data protection reasons. By doing so, you can prevent the AI assistant from ‘gossiping’ confidential information, such as the CEO’s salary, from an Excel spreadsheet. Swisscom supports companies in using Copilot correctly.

From a data‑protection perspective, not providing an official GenAI tool is usually the worse option. Employees who want to use AI will do so anyway. Without an approved solution, they turn to private services, creating so‑called shadow AI, which increases the risk that sensitive data is used for model training and potentially extracted later.

First steps with Microsoft 365 Copilot

Once the data protection tasks have been completed, the aim is to identify the specific use cases for Copilot. The following steps have proven to be helpful. They are worked on in the Swisscom workshops:

• Pilot operation in small groups/teams using Office applications and Microsoft Teams.
  Goal: Develop use cases.
• Knowledge exchange and use of Copilot for all employees, including the use cases that have been developed.
  Goal: Gather experience, witness initial efficiency gains.
• Use business document storage with Copilot.
  Goal: Use of Copilot for analyses, new projects, innovative ideas, etc.

A targeted approach will substantially increase the chances that your company will benefit from the advantages of Microsoft 365 Copilot, as described by the participants in the pilot test mentioned above.