What we learnt going head-to-head with Copilot at our wine tasting event

When we first suggested a wine tasting event in partnership with master sommelier, Tony Seymour of Grinning Goblet Wines, against Microsoft Copilot, there were a few eyebrows raised. Could AI really recommend wines that would stand up against years of human expertise? Could it understand flavour, balance and food pairings well enough to compete? Or would it simply produce a list of safe, predictable choices?

Last Thursday, we put it to the test.

Together with our guests, made up of a mixture of Apex customers and businesses we've recently been speaking with, we enjoyed a five-course taster menu accompanies by ten carefully selected wines (two per course). Five had been chosen by Tony, and five had been selected by Copilot based on the courses served.

Throughout the evening, guests tasted, compared, discussed and scored each pairing with and without food.

While Tony ultimately claimed the overall victory, the real story wasn't who won. It was what we learned along the way.

AI is better than many people expect

Perhaps the biggest surprise of the evening was just how competitive Copilot proved to be.

There is still a perception that AI is little more than an advanced search engine, capable of finding information or letting you know which emails should take priority, but not necessarily making informed recommendations. Yet on several courses, the wine selected by Copilot scored equal or even slightly better than the one chosen by our sommelier.

That wasn't because AI understood wine in the same way a human expert does. Rather, it demonstrated how effectively modern AI tools can analyse information, opinions of experts online, identify patterns an d apply knowledge to a specific challenge. When provided with the right information, Copilot was capable of making recommendations that many wine enthusiasts have been happy to stand behind.

Context changes everything

One of the most fascinating discoveries of the evening came when the food arrived. 

In several cases, opinions completely changed on the wine. A wine that has been the favourite when tased on its own suddenly became a different taste on the palette when paired with food. Meanwhile, another wine that seemed ordinary by itself came alive when paired with the food it has been selected to accompany.

This was a powerful reminder that decisions are rarely made in isolation.

Whether you're selecting a wine, investing in technology or planning future business growth, context matters. The "best" choice is often dependent on the environment, objective and variables surrounding it.

AI can help identify options, but understanding the wider context from a human perspective remains essential.

Human expertise still has a vital role

Tony's overall win reinforced a point that we see regularly when discussing AI with businesses: the most successful use of AI isn't replacing expertise; it's augmenting it.

Years of expertise allowed Tony to understand nuances that AI couldn't fully account for. He could draw upon personal experience, intuition and subtle judgements developed through countless tastings.

Those insights still matter.

In business, we see the same thing. Copilot can help users analyse information faster, draft content, summarise meetings and uncover ideas, but the best outcomes still come when knowledgable people apply their own expertise and judgement to what AI produces. Technology may accelerate the process, but people remain responsible for making the final decision.

The real value is in collaboration

By the end of the evening, it became clear that asking whether AI is better than a human is probably the wrong question. A better question might be: what happens when both AI and a human expert work together?

Many of the guests commented on how impressed they were by the quality of Copilot's recommendations, even when they ultimately preferred Tony's final selections. Rather than proving one approach was superior than the other, the exercise highlighted how the combination of technology and expertise often produces the strongest result.

The most effective businesses aren't looking to replace people with AI.

They're looking at how AI can help talented people spend less time on routine tasks and more time applying the skills, experience and creativity that make the biggest difference.

So, who won?

Ultimately, and rather unsurprisingly to me, Tony took the crown.

But if we're being honest, everyone won. Out guests enjoyed a fantastic evening, discovered some excellent wines and left with a fresh perspective on what AI can and can't do. Copilot didn't replace expertise. It challenged expectations, sparked conversation and proved that when used correctly, AI can be a remarkably capable assistant. And, perhaps, that's the biggest lesson of all.

The future isn't AI vs humans.

It's AI working alongside humans to help them make better decisions, be more productive, and support ideas.