Microsoft 365 Copilot aims to transform meeting prep and productivity

Microsoft has integrated its AI technology into the entirety of the Microsoft 365 productivity suite, with businesses set to experience significant time savings with meetings.

Named 365 Copilot, the AI helper is being added across the Office apps and can cut down time spent writing up meeting notes, combing through inboxes, and drafting talking points.

Copilot was repeatedly branded a “time saver” during a livestreamed demonstration. It can prepare notes and talking points for upcoming client meetings, draft documents in Word, analyse and visualise dense Excel spreadsheets, and pick out key emails from a busy inbox.

Among the most notable features was Copilot's ability to prepare documents ahead of meetings. It works by reading data from across the 365 suite to provide recommended reading and upcoming deadlines with which staff can familiarise themselves ahead of time.

Copilot’s Business Chat feature can also take data from meeting minutes, emails, calendars, and chats to provide a summary of interactions with a specific individual, complete with containing citations and contact cards on key customers.

The AI tool is also capable of summarising events in a live meeting, giving commentary on how a group feels about the contents so far, and providing suggestions on talking points that still need to be touched on.

As a call comes to a close, Copilot can generate a list of next steps saveable directly into the user’s chosen Customer Relationship Management (CRM).

For further improvements to productivity, this relationship goes both ways as Copilot can draw on CRM records to contextualise emails, meeting plans, and other automatically generated communications.

Microsoft Copilot 365: Email integration

Copilot harnesses user context to cut down time spent manually drawing together data or processing information, even in tasks that have rarely been automated in the past. An example of this is the tool’s email ‘catchup’ feature, which rounds up emails it has identified as particularly urgent.

“We all struggle with the pull of the inbox,” said Sumit Chauhan, CVP of the Office product group at Microsoft.

“Sometimes it keeps us from focusing on our most important work. But at the same time, you do have important emails to respond to. Copilot separates the signal from the noise and gives you hours of time back. Copilot can help you triage your inbox, highlighting the most important emails to prioritise.”

Copilot will also be available on the Outlook app, and can also summarise long email threads, and convert documents such as a sales spreadsheet into digestible texts and graphs within the email client. All text generated in this manner is fully-editable, and users can change the tone or brevity of emails with one click.

Microsoft Copilot 365: Smart documents

Selecting and converting documents is a core function of Copilot, and Microsoft singled out its ability to helpfully draw data from across a range of locations in real time as one reason the AI tool could entirely change the way people work.

Copilot is powered by a processing and orchestration engine it is calling ‘The Copilot System’, which combines 365 apps, Microsoft Graph, and a dedicated large language model (LLM).

Microsoft Graph is the API developer platform for Microsoft 365 - it comprises content and context from files, chats, calendar, and emails. Microsoft stated that Copilot draws on all of this to ensure that results are as relevant and contextualised to each business as possible, and that the process is carried out in a secure way to preserve privacy and maintain compliance.

The firm also specified that Copilot uses GPT-4, OpenAI’s powerful multimodal model announced this week, for the LLM portion of the ‘Copilot System’. The LLM trained on a “large but limited” amount of a business’ data, and the firm stressed that tenant data and prompts are not used for training purposes.

In bringing together apps, content, and context from across a customer's suite, Microsoft has set its sights on full intelligent automation of tasks.

This is reflected in the shared name scheme of its recent AI productivity announcements, with GitHub Copilot, 'AI copilot' for Edge, and now Microsoft 365 Copilot cementing the degree to which the company is set on making helpful AI a brand experience rather than limiting it to a particular software.

Microsoft Copilot 365: Extra time-saving powers

The cross-discipline time-saving power of Copilot was also emphasised in a demonstration showing how it can make Microsoft’s Power Automate more accessible.

Users can request an automated workflow using natural language and Copilot will produce a diagram to explain the workflow and allow user edits.

Microsoft showed an example in which a developer requested Power Automate to create a private Teams channel whenever a specific customer sent an urgent support ticket.

After its initial summary of the proposed workflow, Copilot was able to take feedback from the developer and ensure that when the event was triggered. Copilot should also summarise the support ticket and tag relevant departments that can help.

"Although people may initially resist the idea of using AI, the ability to do things like summarise long documents and prioritise your email inbox will be an absolute godsend for most people struggling to cope with the ever-increasing volume of information we have to get through every day,” said Dr Andrew Rogoyski, director of innovation and partnerships at the Surrey Institute for People-Centred AI.

"Writers, copyrighters, screenwriters, and journalists will have to integrate these new tools if they want to match the productivity of their competitors.

"My guess is that the next step will be for Microsoft and others to offer enterprise bolt-ons to Copilot that incorporates all the corporate knowledge in a particular company, thus allowing customised versions of Copilot to provide bespoke help to individual companies.

"This is yet another blow for Google as companies who have invested in the Google Workspace, which grew massively during the pandemic as users shifted to the cloud, may be asking themselves if they backed the right horse."