I suspect I wasn’t the only one who watched Microsoft’s livestream with my mouth hanging open. Last Wednesday (03/16/2023), the livestream took place under the promising title «
The Future of Work: Reinventing Productivity with AI
» took place. And then Satya Nadella and Jared Spataro dropped the bombshell: M365 Copilot. They thus more than lived up to the title of the event. Because for me personally, this is definitely a gamechanger for our daily work. But let’s go through the announcements together, step by step:

  1. Why is M365 Copilot needed?
  2. What is the copilot approach? & What are Large Language Models?
  3. How does the copilot work? incl. Preview in the individual apps
  4. My personal thoughts about M365 Copilot
  5. What’s next?
  6. I am interested in your opinion!
  7. Resources / Sources

Why is M365 Copilot needed?

Nowadays, we spend a lot of time doing work we don’t enjoy as much instead of focusing on what’s important. The work we are best at and enjoy the most. Wouldn’t it be great if you could just say what you needed and someone would do that work for you? I think so too! You’ll see, M365 Copilot does just that for you. Microsoft has only recently made a significant breakthrough with the introduction of

Teams Premium

 showed how valuable AI support can be in everyday life.

So, you can say that Microsoft365 Copilot will help you to be even more productive, creative and better at work. You can get a first impression of Copilot in the Youtube video from Microsoft.

What is the copilot approach?

Microsoft has also successfully implemented the Copilot approach or made further announcements in other areas. How to use
Github Copilot
 is already live and initial survey results show impressive numbers:

  • 88% reported being more productive
  • 77% stated that it took them less time to search or find information
  • 74% said that they were able to focus more on the work they enjoy

And that is precisely the goal of the copilot approach. To be able to use the invested time even better and more efficiently. To have your personal co-pilot at hand at all times, so to speak, who supports you directly. At the latest since the chat GPT, the majority has once again become aware of how powerful artificial intelligence can be and, above all, how useful it can be in our everyday working lives.

With M365 Copilot, the so-called Large Language Models (LLM) make their way directly into our everyday lives. Speaking of which, what are Large Language Models? This can be explained mathematically. Like I did e.g. in my computer science studies in the lectures «Artificial Intelligence 1 and 2», or you just ask Chat-GPT 🙂 (If only I’d had that in my studies…).

so I asked Chat-GPT – Pascal: «
Explain Large Language Models
«

CHATGPT: Large Language Models (LLMs) are artificial intelligence systems trained to understand and generate natural language. These models use deep learning techniques such as neural networks to process enormous amounts of text data, automatically detecting patterns and relationships in the language.

An LLM consists of millions or even billions of parameters and, due to its large size, is capable of handling a wide range of natural language processing tasks. These include text comprehension, text generation, machine translation, question-answering, text summarization, chatbot communication, and more.

LLMs are trained by training on large datasets of human texts such as books, articles, Wikipedia entries, and more. Training is done through a process called supervised learning, where the model is trained with a large amount of sample text to enable it to recognize patterns and rules in the language.

A well-known example of an LLM is OpenAI’s GPT-3, which is one of the largest and most advanced LLMs developed to date, with 175 billion parameters.

Pascal: «
Explain it to a child
«

CHATGPT: Large Language Models (LLMs) are like super smart robots that help us understand and speak language better. Just like when you read a big stack of books and learn lots of new words and ideas as a result, an LLM reads millions of books, articles, and other texts to learn how to use language.

Once the LLM has learned enough, it can help us do many different things with language, such as telling stories, answering questions, or even speaking in different languages. The bigger the LLM is, the more it can learn and the smarter it becomes.

All clear! So now we know: LLMs read very, very much and can also remember what they read and don’t forget anything (still kind of cool) and this technology is now coupled with the possibilities of Microsoft 365. That’s because M365 Copilot combines the power of LLMs with data from Microsoft Graph – calendars, emails, chat messages, documents, meetings and much, much more. The whole thing is processed accordingly and since LLMs are in use, this knowledge can be retrieved in natural language.

How does the copilot work?

As mentioned earlier, the big strength is the combination of LLMs and enterprise data via Microsoft Graph. This means that the co-pilot will find its place at various levels in everyday work. Currently Microsoft is talking about two areas

  • Integrated with Office apps (Word, Excel, Outlook, Teams, etc.)
  • Business chat

It is worth mentioning that the first use cases for integration in Microsoft Viva have already been presented. Let’s take a look at the different deployment scenarios together.

Microsoft Word

Have you also had very long documents in front of you where a quick summary would have been helpful? Have you ever had writer’s block and felt like every word you typed didn’t fit? This is exactly the kind of moment when Copilot in Word can help you. Because Copilot can enrich your existing Word document with additional content, summarize texts or revise sections – not to mention that you can even set the tone, which further supports your productivity. And as we have learned, you can easily put your requirements in writing, e.g:

  • Design a two-page project proposal based on data from [Wähle ein Dokument aus] and [wähle eine Excel-Datei aus]
  • Make the third paragraph more concise. Change the tone of the document to make it more casual
  • Create a one-page draft based on this rough outline

Microsoft Excel

Have you ever had an Excel file in front of you and taken a long time to understand the data and figure out which column means what? Here, too, the co-pilot will help you with advice and action.

  • Give me a breakdown of sales by type and channel and include a table
  • Project the effects of [change a variable] and create a diagram to illustrate it
  • Model how a change in the growth rate for [Variable] would affect my gross margin

See for yourself:

Microsoft PowerPoint

What sells and presents best? Exactly – stories! But we all know that we have our creative phases and, unfortunately, often our uncreative phases. In addition, not everyone enjoys creating PowerPoint presentations equally. Copilot in PowerPoint will help you conjure up a memorable and stunning presentation in no time. (I think the word conjure really sums it up well).

  • Create a five slide presentation based on [Word-Dokuments] and include relevant stock photos
  • Summarize the presentation in a three-part summary
  • Format these three bullet points into three columns with one image each

Microsoft Outlook

Have you ever been added to the CC after a real mail ping-pong? Wouldn’t it be great to get a brief summary of what was written and who has what opinion? and whose mailbox is about to explode after the vacation? Wouldn’t it be great to have someone who goes through all the mail and marks the most important ones? and yes – also here comes the copilot

  • Summarize the emails which I missed last week. Mark all important mails
  • Write a response thanking me and asking for more details on points 1 and 2; brevity this draft and change the tone to business
  • invite everyone to a «lunch and learn» about the new product launch next Thursday at 12 noon. Mention that lunch will be provided

Microsoft Teams

Copilot, of course, doesn’t stop at our most important collaboration tool, teams. On the contrary, have you ever been late to a meeting? This is especially annoying when you don’t know in detail what has already been discussed and who holds what opinion. Or after a heated and good discussion, it would be helpful to list all the mentioned advantages and disadvantages again. This is also where Copilot comes into play:

  • Summarize what I missed in the session. What points have been addressed so far? on which topics do we disagree?
  • Create a table of pros and cons for [diskutiertes Thema]. What else should we consider before making a decision?
  • What decisions have been made and what next steps are proposed?

Business chat

And then there was another announcement from Microsoft: Business Chat. A completely new experience that works with all data and apps. This allows information from different sources to be brought together so that all team members are on the same page and moving forward together. With the goal of spending less time with tools and more time doing effective work. With the following examples:

  • Summarize the chats, emails and documents on the [Kunde] escalation that took place last night
  • What is the next milestone at [Projekt]. Were any risks identified? Help me brainstorm a list of possible remedies.
  • Write me a new planning overview in the style of [Datei A], which includes the planning schedule from [Datei B] and incorporates the project list from the email from [Person]

Microsoft Viva (engage)

Copilot will be integrated with Viva Engage, among others. It helps to create even better posts and communicate even more clearly. In addition, Copilot allows for an even better ‹Answers› experience

Power Platform

Of course, the copilot does not stop at Modern Work, but is integrated directly into the Power Platform. The Microsoft Power Platform is already a powerful toolset that is still far too underutilized. But wouldn’t it be awesome if you could describe a mobile app and the AI would build it for you? Deliver here too
Copilot in Power Apps
 and

Copilot in Power Automate

 provide the answer! Another gamechanger moment when we think about developing solutions with the Power Platform, see for yourself:

My personal thoughts about M365 Copilot

During the livestream, I was actually just thinking «WOW, what a gamechanger for our everyday lives». The longer I think about it, the stronger this thought becomes. Especially if you think about the use cases even further and include elements such as
Viva Topics
 it becomes clear how powerful the M365 platform can be for companies and their employees. The freed-up capacity enables a new way of organizing everyday life. All information and the knowledge of where it is located on each Microsoft tenant can be made accessible. Even if we think of things like 4-day work weeks, this approach sounds very reasonable. With less time you can achieve the same or even more. Things that seem annoying are done in a few sentences, and all in a personal context. Of course, the responsibility of artificial intelligence also plays a major role. Microsoft has made a clear commitment to «
Responsible AI
«. Read also a
published document of the Microsoft
. The new technologies also reinforce the chosen compliance/security policies, so that each employee can only see the information to which he or she is entitled.

As always with such technologies, they must of course prove themselves in everyday use. The validation of the information must also be in the foreground. Just because artificial intelligence is behind it doesn’t mean it’s correct or reliable in every case. The big advantage, however, as read at the beginning, is that AI can read very, very much and at the same time forgets nothing. In addition, it is constantly learning, i.e. the longer it is in use, the more accurate it becomes. This is also confirmed by a note in Microsoft’s blog post:

Finally, we have worked to design an experience that prioritizes human agency and puts the user in control. This includes noting limitations, providing links to sources, and prompting users to review, fact-check, and fine-tune content based on their own knowledge and judgment.

In summary, it is indeed a gamechanger for me and I can’t wait to test it in my environment and benefit from this productivity gain. At this point, however, it should also be mentioned that companies that are not yet using M365 should jump on the bandwagon now at the latest in order to be able to benefit from the countless advantages in the medium term.

What’s next?

Very many, of course, want to try this out directly and profit from it. Microsoft is currently testing the Microsoft 365 Copilot with 20 customers, 8 of which are from the Fortune 500. The preview will be further expanded in the coming months. Of course, it must also be observed in what form IT administrators can exert influence and how such a powerful tool can be integrated into corporate governance. First insights can be found under the following article:
Administration of Microsoft 365 in the new era of AI

As seen in this article, Copilot doesn’t stop at teams, but will make its way into all areas. The Power Platform will also be equipped with Copilot functionalities, as will the Dynamics365 world. This gives us perfect support in all areas for our daily work.

We will also learn more about the price in the coming months. But when you look at the efficiency and productivity gains, I can imagine that some companies will invest in such licenses.

In any case, the development of Microsoft 365 is very exciting and I am very happy to be able to accompany you on this journey.

I am interested in your opinion

Now, of course, I’m wondering what you think? What do you think of this announcement? Where do you see the biggest advantages, but also dangers/risks? Let me and the other readers in on it. I’m eager to hear your assessments.

Resources / Sources

The article and the images and GIFS used were written based on various sources:

Folge mir auf LinkedIn