Implementing Microsoft 365 Copilot

Sean Wells at Atos outlines his top three lessons for businesses looking to adopt Microsoft 365 Copilot

 

For the last 9 months, I’ve been a Microsoft 365 Copilot early adopter for Atos, a global leader in digital transformation. Now, I will be passing on my learnings and advice, hoping to help and guide businesses that are considering adopting the tool.   

 

When I first heard that “Microsoft 365 Copilot is an AI-powered assistant integrated into Microsoft 365 apps like Word, Excel, PowerPoint, Outlook, and Teams,” I wasn’t exactly sure what it was or, more to the point, how I’d use it. 

 

Indeed, I wondered, with all the initial hype, whether Copilot was a solution looking for a problem. But 9 months on, it’s fully embedded into my daily work activities.

 

To help businesses keen to integrate this technology into their daily operations, I’ll share my top three insights from my experience as a Microsoft 365 Copilot early adopter, along with some practical guidance.

 

1. Don’t over think it

Don’t overthink what your use cases will be at the outset. Go ahead and experiment, at a personal and corporate level, before making the next big step.

 

Some organisations struggle to get a Copilot launch signed off without prior agreement of tangible use cases and a business case. That’s a chicken and egg situation. Copilot is an emerging technology, and its true potential comes from practical experience. Excel just had its 40th birthday, and few could do their job without it nowadays. Copilot, or future generations of it, will be just the same.

 

My advice is to go ahead and experiment. Invest in a focused launch with a target set of users: that could be within a function or a cross-section of users from across the organisation. Take regular checks on who, why, and how Copilot is being used. Share experiences and learnings between users. Then take stock. You’ll probably find users are commonly using it for:

• Drafting and editing documents and generating content in Word and PowerPoint.
• Summarising long email threads in Outlook or complex data in Excel.
• Getting insights and recommendations for data in Excel.
• Summarising meetings and providing action items in Teams. 

These personal productivity benefits feel very real to a regular user but can be hard to quantify in terms of a tangible financial return. 

 

Get your launch right, and use cases more integral to the business, with clearer financial returns, can emerge, for example in:

• Customer service: Responding to customer enquiries and managing support tickets.
• Marketing: creating materials, drafting social media posts, and analysing campaign performance.
• Finance: reconciling data, generating financial reports, and optimising financial operations.
• Human resources: Improving onboarding processes, managing employee data, and streamlining recruitment.
• Project management: Streamlining project planning and tracking in Planner and Project. 

Those use cases tend not to emerge by sitting in a room and thinking where and how Copilot could be applied. Experience to date suggests it’s better to allow users to experiment for a period to identify potential areas, and then take stock to really focus on those that have the biggest impact, and really make those a success in your next big step.

 

2. Take advantage of the training and user groups

Copilot is very intuitive but take advantage of training that’s available and user groups too to maximise its impact.

 

I found Copilot hugely intuitive and easy to begin to use. That’s a good thing, and quickly I got to the stage of using it daily for emails, research, prioritising my day, and getting a summary on a particular issue. 

 

By just typing “Tell me what dealings my organisation has had with company X,” Copilot gave me a summary of my emails and Teams messages together with documents that I had either forgotten about or never even knew existed. I want to stress here that we have tight data governance and security controls to ensure I only gain access to what I’m entitled to. See point three below for more on this critically important lesson!

 

I now wish I’d taken earlier advantage of training, drop-in sessions, and user groups. When I eventually started using them, a whole new set of capabilities became apparent and immediately useful.

 

The lesson here, therefore, is two-fold: if you’re a Copilot user, then by all means jump straight in to see what you think, but also get some training along the way; and if you’re responsible for implementing Copilot in your business, then make training readily available. Set up user groups too, to share experiences and flush out some of the business-critical benefit areas listed above, e.g., in customer service, finance, or marketing.

 

Part of that training must include education on responsible AI practices, to ensure that Copilot’s AI is used in a way that aligns with your organisation’s values. Which brings me to the following…

 

3. Prepare for possible problems

It can go wrong. Ensure your organisation does the groundwork in advance.

 

I was lucky. Prior to launch, my employer took care of the fundamentals needed around data readiness, security, compliance, and licensing. That also included confirmation of the boundaries, principles, and policies of where and how Copilot would be used, e.g., what data Copilot data sources and integrations were acceptable, with controls established to prevent access to those that were not. It was only when I began talking to other organisations that I learned of horror stories and pitfalls some are seeing due to a lack of preparedness. 

 

For example, in one company, Copilot began discovering and making visible a set of payroll documents that really should have been kept confidential, what’s called content oversharing. Lack of proper data governance and security were the underlying problems; it was just that Copilot exposed the issues. 

 

Copilot uses existing permissions and policies to deliver the most relevant information. This means it is important to have strong content management practices in the first place. That organisation should have ensured their data was properly categorised, with associated secure access rights, before launching Copilot in their business.

 

Where next?

For me? I’m hooked on Copilot. My next step is to take advantage of more of the training to help me do more and be an ambassador with my team and peers.

 

We’re also helping our clients with Copilot in many and varied ways: in making their first steps or expanding usage to wider users; driving out tangible use cases and business cases; and ensuring they’ve the right data security and governance before launch.

 

---

 

Sean Wells is Private Sector Sales Director at Atos

 

Main image courtesy of iStockPhoto.com and Kenneth Cheung