Who will benefit the most from Microsoft 365 Copilot?

I was recently asked to do a briefing on Microsoft 365 Copilot. The timing was a little challenging because Copilot is currently only available to a select group of preview customers and most of the content, we see is marketing material and planned demonstrations.  

Like many people I have experienced ChatGPT and other similar AI tools and seen the potential as well as the challenges these tools bring. AI in the work context is new and brings with it many challenges for organisations:

  • What use cases are OK and what are not?
  • What are the risks with our own content?
  • How do we train our people?
  • How do we realise the benefits?
  • Are we going to be left behind?
  • And so on…

Copilot brings AI capabilities into the tools we use every day. It leverages a Large Language Model (LLM) along with our own content and the Microsoft Graph to create a unique AI experience in the context of the work you are doing. It has enormous potential to improve productivity for those repetitive and time-consuming tasks we find ourselves doing frequently.

The Microsoft Work Trends Index 2023 report highlights the problem. We spend more than half our workday communicating in Teams, Outlook and in meetings (57%). The remained of our day is spent creating or actioning the items (43%). You read that right, less than half our day is spent doing!

In my own personal experience, I often run out of time to fully read communications in Outlook or Teams. I often read the last message and not the entire thread because it comes at the cost of doing other things. This creates a problem, the illusion of communicating. Someone sends you a message with a task buried on the 4th paragraph, that you miss. The communication is ineffective, and things get missed. It also leads to additional communication time that may not have been necessary.

Perhaps the biggest part of my job is writing. I produce reports, respond to correspondence, and produce a variety of other content. Much like this blog, the hardest part is getting started and producing something semi-coherent. I learnt when writing a book that the biggest challenge is time. What if we could spend less time on the getting started and more time on the finishing?

MIT did some interesting research on the use of AI by researchers writing papers. They found that not only did authors on average save 40% of the time needed to write a paper, but they also produced something that was 18% better quality than papers produced without the support of AI.

Boston Consulting Group (BCG) research found that the consultants who are average performers had much more to gain from using AI that those who were already high performers. Why? Average performers take longer to do things and their output is usually lower quality, so the ability to improve their performance is greater. Not only that but it takes up less of other people’s time because there is less re-writing of content. Think of a lower-level consultant producing something that is reviewed by a senior (expensive) consultant before sending. How much time does the senior consultant use, and can AI help save some?

BCG also found that in some cases AI had a negative impact on performance. People performed 20% worse if ChatGPT couldn’t answer the question or gave poor responses. In other words, they would have been fast if they didn’t use AI to assist. Training will be required, don’t just ask IT to provide the tools and walk away.

I frequently read comments about software developer’s using AI tools such as GitHub Copilot where an experienced developer will say “I don’t see the need for these tools”, well I think BCG nailed it. The point is, we aren’t all highly experienced consultants (or developers), we aren’t all highly productive writers of content or code and most of us what to be better.

In conclusion, I think we should look at the average performers in our teams when we look to see where AI would benefit us. What do they produce that others need to review? How do we reduce the load on the people upstream from their initial drafting? Lifting the average performance of your team will make the boat go faster!


Discover more from SharePoint Moments

Subscribe to get the latest posts sent to your email.

One comment

Leave a reply to Bing Chat Enterprise – SharePoint lightbulb moments Cancel reply