Now boarding: What SMEs need to know before using Microsoft’s Copilot

It’s become a bit of a status symbol in the technology world — have you got Copilot?

The answer is almost always: not yet.

The scarcity of access to Microsoft’s collaboration with OpenAI means the vast majority of Australian businesses don’t yet know what Copilot will do for them, whether it is worth it or when they might be able to see it for themselves.

Yet if early research from Microsoft proves correct, it will become a tool that could end up as a differentiator and competitive advantage for a wide range of businesses. The first Australian companies to get access to Copilot in Microsoft 365 only started their pilot programs back in September, with just a few very large organisations tapped to join.

That list included some law firms, major insurers, and superannuation firms.

By November, Microsoft broadened access to Australian Microsoft Enterprise customers but limited it to those with more than 300 seats under a single tenant. The company says it plans to reduce the threshold for access and has launched a sort of Copilot-lite in Windows 11, but it will be a while before most organisations can really assess the usefulness of having large language models embedded across tools like Outlook, PowerPoint, Word and Excel.

So what do we know about Copilot, based on those who have tried it?

While based on a very small cohort, Microsoft’s research into the earliest users of Copilot has shown enthusiasm for its ability to make work easier and faster.

The vast majority said it made it easier to get started on a first draft of content (87%), made it easier to take action after meetings (84%), and saved time by finding information across files (75%).

The actual time savings are still relatively low — 14 minutes a day on average, with one in five saving 30 minutes or more — but that could grow as more users join the program and start to work in ways that Copilot better supports.

Specific groups also benefited more than others. People using Copilot across sales and CRM estimated a 90-minute saving a week.

Findings on the quality of work were also interesting. When emails were shown to a blind panel who were asked to rate the quality of the work, not knowing whether the author was human-only or human-plus Copilot, the AI-supported work was ranked 18% clearer and 19% more concise.

Perhaps the most important finding for business, though, is that once tried, more than three-quarters of people don’t want to give it up.

In fact, 88% said they would rather have Copilot than a free lunch each month at work. One in three said access to Copilot would influence their choice of employer.

At Pitcher Partners we are already trialling Copilot and like other early adopters, we are experimenting and putting it to the test with several use cases. While we are early on this journey, we are optimistic that it will help us better assist our clients and better support our people to be more efficient.

What might it mean for companies beginning to look at their Copilot options?

The first is that this should be a consideration for those businesses with large Microsoft tenants, who already have well-adopted use of products like Teams or Dynamics, and who use Microsoft file storage on SharePoint or similar platforms.

For these companies, the addition of Copilot could create incremental gains right across their processes which could allow resources to be focused on more productive activities.

A key challenge though will be cost. For a large enterprise customer, Copilot could be tens of thousands of dollars more each month.

What about those who are less enmeshed in the Microsoft world? For these companies, the decision will be more challenging.

If gaining value from Copilot also means abandoning existing, tried-and-tested platforms, it might be better to see how those technologies plan to implement similar generative AI approaches.

As an example, companies that use Salesforce instead of Dynamics can access a range of AI options across its CRM, with its Einstein product built on OpenAI’s ChatGPT.

AI options are also rapidly being implemented in ERP software like Oracle and SAP, finance software like Xero, and marketing and design software like Adobe Suite or Canva.

But there might also be an argument for reviewing the spread of data and information across platforms.

MuleSoft, in its annual connectivity benchmark, estimates the average number of apps for an enterprise ecosystem blew out to 1061 in 2023, up from 976 a year earlier, with under a third of those applications integrated.

The case to reunite the tools and processes used across your business and improve the Microsoft experience with Copilot could be a compelling one for companies struggling to contain app proliferation.

Right now, for many businesses, this decision is moot as to whether to unify data in order to access AI through Copilot or focus instead on an AI-enhanced software stack, given access to the Microsoft product is so restricted.

But as 2024 rolls around this will be a key question that companies and their IT leaders will need to assess.

Martin Koval is the client director of digital enablement at Pitcher Partners.