The risks and benefits of using Copilot

Sebastian Dewhurst at EASA asks whether Excel’s Copilot is a revolution or a risk

 

The AI revolution isn’t just on its way; it’s already elbowing its way into our daily workflows, promising to free us from the tedium of spreadsheets. Excel’s Copilot is a prime example, aiming to simplify data analysis and improve decision-making through natural language.

 

However, while Copilot offers significant benefits, we must also recognise its limitations. While Copilot offers the allure of rapid analysis and streamlined decision-making, it also carries the potential for hidden pitfalls. Can we trust its interpretations? What about the security of our sensitive data? And what happens when we become too reliant on its automated insights? 

 

The reality is that Copilot is a double-edged sword. To unlock its true potential, we need to consider both its power and its pitfalls. Let’s dive into how to leverage Copilot’s strengths while avoiding the traps that can trip you up — drawing on real-life case studies to show you exactly what works and what doesn’t.

 

 

Pitfalls of relying on Copilot

While Excel’s Copilot offers tempting efficiency gains, businesses must be acutely aware of the potential pitfalls associated with its implementation.

 

Compliance blind spots

One issue is that Copilot doesn’t enforce compliance. It’s not designed to handle data integrity, version control, or industry regulations. This is a big deal, especially in sectors like finance or healthcare, where strict adherence to rules is non-negotiable.

 

Take Metro Bank, for example. The British commercial bank was hit with a £16.7 million fine because its system, which was supposed to detect money laundering, was completely flawed. Between 2016 and 2020, the bank didn’t monitor over 60 million transactions worth billions of pounds due to technical issues. This happened even though employees raised concerns. So, despite all the automation, the lack of proper oversight led to the eyewatering fine.

 

The point is, if you’re using Copilot to generate reports without keeping an eye on the compliance side, your company could be exposed to serious risks. You can’t just assume the AI will handle it all — strong controls are necessary.

 

The human element still matters

Copilot is a powerful tool, but it’s not a substitute for human judgment. It can crunch numbers and generate insights, but it can’t understand context or make nuanced decisions. If you blindly follow its suggestions without critical evaluation, you’re setting yourself up for potential errors.

 

Here’s a good example: As Reuters reports, a Texas lawyer recently got fined after using an AI tool to help with legal research. The AI tool generated fake citations, and because the lawyer didn’t double-check the results, the lawyer ended up submitting bogus info in court.

 

Relying too much on AI, like following an AI-generated sales forecast without considering external factors or market trends, could lead to a strategy that just doesn’t work. Remember, AI is there to support your decisions, not make them for you. It’s about using its insights to inform your choices, not replacing your expertise.

 

Data security risks

Data security is a big deal, and Copilot could open up potential security risks. If you’re feeding it sensitive information, you need to think about how it’s handled.

 

Coinbase, the cryptocurrency exchange, got slapped with a $50 million fine for serious compliance failures, including exposing sensitive customer data. The company had some major vulnerabilities in its system, which put customers at risk for things like fraud and money laundering.

 

So, if you’re using Copilot with confidential data like financial reports or customer information, you have to be aware of the risks. Are you comfortable with the possibility of sensitive data being exposed? It’s crucial to have strong data security protocols in place and to make sure employees know how to protect sensitive information. Don’t take data security lightly, especially when dealing with AI systems that process large amounts of information.

 

 

The benefits of Excel’s Copilot

Still, despite a few hurdles, the positive impact Copilot can have on workflow and data management is undeniable:

 

It makes things faster

Copilot significantly reduces the time and effort required for routine data manipulation. It automates analysis, pulls out insights, and handles repetitive, time-consuming tasks we all dread. Essentially, it lets us focus on the work that actually matters.

 

It helps you understand your data

Copilot suggests formulas, points out trends you might miss, and even makes the data look presentable with charts. It’s like having someone help you connect the dots. You get those lightbulb moments faster, and it’s way easier to explain your findings to others. It’s not just about numbers; it’s about making sense of them.

 

It takes the pain out of formulas

Copilot lets you just tell it what you want, and it works out the formula for you. No more endless searching for the right function or making mistakes. This is a game-changer for anyone who’s not an Excel expert. This feature empowers users to perform sophisticated analyses with greater confidence and accuracy, leading to more reliable and informed outcomes.

 

 

How Copilot is evolving

Excel’s Copilot is a big leap forward in AI-assisted data analysis, but it’s still a tool that works best when paired with human oversight and solid data management practices. We shouldn’t see Copilot as a one-stop solution. Instead, view it as a tool that’s evolving and, when integrated with enterprise applications, can help businesses simplify complex data tasks, all while keeping control, compliance, and security in check.

 

More automation

As AI keeps improving, Copilot is expected to take on more advanced workflows. Right now, it helps with tasks like generating formulas, visualising data, and recognising patterns. But future versions will likely go further, such as: 

• Automating financial and operational reporting with little user input.
• Using AI to model scenarios and support decision-making, like forecasting and risk assessments.
• Integrating with external business intelligence tools to create fully automated reporting systems.

However, with more automation, businesses still need to keep an eye on things to make sure AI-driven insights stay accurate, reliable, and in line with company goals. 

 

Better natural language processing

One of the most exciting advancements in Excel’s AI tools is the improvement of natural language processing (NLP). In the future, Copilot will likely feature: 

• More intuitive and conversational interactions, letting users request detailed analyses in plain language.
• AI-driven explanations of complex formulas and logic, so non-experts can better understand advanced calculations.
• Smarter suggestions that take into account past data trends and company-specific details. 

These upgrades will make Excel’s AI tools easier for more people to use, cutting down on the learning curve and encouraging wider adoption. 

 

Stronger human-AI collaboration

Even though AI tools like Copilot can process vast amounts of data quickly, human expertise is still essential for understanding context and making strategic decisions. We see a future where: 

• Copilot acts more like its namesake, a “co-pilot” than a replacement for skilled analysts, guiding users through data exploration without making final decisions.
• AI helps with data validation by pointing out inconsistencies or issues for human review instead of assuming everything is correct.
• Collaboration features improve, letting multiple team members interact with AI-driven insights in real-time, leading to better teamwork and sharing of data across departments.

 

Finding a balance

Excel’s Copilot has massive potential, but its true value comes from how businesses use it as part of their overall data strategy.

 

As AI keeps evolving, organisations need to find the right balance between automation and human expertise, ensuring their data processes stay accurate, compliant, and adaptable as their needs change.

 

---

 

Sebastian Dewhurst is Founder at EASA

 

Main image courtesy of iStockPhoto.com and sankai