There is a quiet revolution happening across British business. Not in the press releases, not in the glossy trade show stands, but in the back offices of distributors, the Excel spreadsheets of wine merchants, the inboxes of sales teams who spend half their working lives doing things that have nothing to do with selling wine. Artificial intelligence is transforming how businesses operate, how they compete, and increasingly, how they survive. However, the drinks industry, for all its charm and tradition, is at risk of being left behind.
That was the uncomfortable but galvanising truth at the heart of a panel session held at this year's London Wine Fair. The final session of the show, titled 'AI: Friend or Foe?', brought together five people who work at the coalface of technology and drinks. They are operators, consultants and founders who are already using AI not as a novelty but as critical infrastructure. What emerged over the course of the hour was not a tech pitch or a warning about robots stealing jobs. It was something more nuanced and more urgent: an honest assessment of where the industry is, what it could become and what it risks losing if it continues to drift.
The here and now
The panelist included Mitch Fowler, chief executive officer and co-founder of Fero (formerly Ferovinum), James Grant, co-founder of consultancy firm iwantmore.ai, Paul Cartwright, chief technology officer at TVision, Laura Rosenberger, an AI consultant and NIck Martin, chief executive officer and co-founder at Wine Owners. The session started with a bit of revelation from the moderator, Richard Siddle, editor at The Buyer, before a word of the debate had even been spoken. He confessed he had invited the top ten drinks distributors in the UK to participate. Not one said yes. Some admitted they weren't using AI at a level they felt comfortable sharing publicly. Others simply declined.
"We're all in this big wine fair here today and actually how many companies here are actually using AI at all?" — Richard Siddle, The Buyer
The answer, if the room was honest with itself, was laid bare: not many. And those who are using it are mostly using it in the most basic way possible – a free ChatGPT account, perhaps, generating a bit of copy or answering a quick question. There is, as several panellists noted, a pervasive anxiety about AI in the industry. Employers worry their staff will think their jobs are threatened. Staff worry their employers are right. Leaders who've never used the tools don't know where to begin and so they don't begin at all.
Long term, this response won’t be sustainable, but it is also an entirely understandable one. Understanding why the drinks industry feels this way is the first step to moving past it.
A relationship industry in a data world
The drinks trade is, at its heart, a relationship business. It always has been. Whether you are a Bordeaux négociant, a UK distributor with thirty years of contacts or a wine merchant who knows every customer's palate, the currency of this industry is human connection. Trust, taste, personality, presence. These things cannot be automated and nobody on the panel suggested they should be.
But there is also the reality that most people who work on the relationship-facing side of the industry spend an alarming amount of their time not building relationships. They are logging into systems. Updating spreadsheets. Chasing invoices. Reconciling stock. They are drowning in administrative work that, in another era, simply could not be avoided. In this era, much of it can.
"Salespeople do admin. People who are spending far too long logging into multiple systems to try and update stuff or check stuff — if that can be dealt with, then everyone's happier in their work too." — Nick Martin, Wine Owners
The opportunity AI presents to the drinks industry is not to replace the people who look after those relationships or know the range inside and out.. The opportunity is to give those people back the hours that have been stolen from them by process and admin.
The perception problem
Ahead of understanding what AI can do for a business, people need to understand how to use it, but using it incorrectly can be damaging to those hard won relationships.
"Most people still use AI as just a question-answer thing," said James Grant, co-founder of consultancy firm Iwantmore.ai. "I'm going to ask ChatGPT a specific question. I'm just going to get back a bit of text. The number of emails you see coming out where people have just used AI to generate an email, just sent it off without any thought as to how that's perceived on the other end is horrific. That's where you're losing that credibility, you're diminishing those relationships. AI can help with this and do it well, but people do need to understand how to use these tools properly."
This is perhaps the single most damaging misuse of AI in a relationship-driven industry: using it to hollow out the very thing that gives the industry its value. A mass-produced AI email, stripped of personality and context, sent to a buyer when a business has spent years cultivating that relationship, is not just not a productive use of AI, but a potentially damaging one. Not because the tool is bad, but because it is being misused by someone who doesn't understand it.
"It's not AI that will take your job. It's the person who's leveraged it properly. And it's the same with businesses." — Mitch Fowler, Fero
Free or paid? The first decision every business must make
One of the most practical exchanges of the session came when an audience member asked what is the difference between free ChatGPT and paid AI tools — and when should a business make the switch?
Paul Cartwright, Chief Technology Officer at Tvision, gave a clear answer: "Fundamentally, free ChatGPT and paid Claude, or Copilot, or whatever — they are the same thing, a large language model. What you're paying for is two key things. One, you're no longer the product. When you use the free ones, you do not know where your data is going. Two, as you move into using AI in business, you're moving beyond a chatbox. Taking that brain out of the chatbox and using it in business processes — putting harnesses around it — is where the real value lies."
Nick Martin was even more direct: "If you're a business, don't be using the free version. That's probably the simplest way to look at it, because you're giving it your data and you don't want to be doing that."
The good news is that the paid versions are extraordinarily affordable by any standard of business technology investment. Mitch Fowler made the point bluntly: "If you can't find thirty dollars of value out of a monthly subscription to Claude, or Copilot, or Gemini, whatever you're using, you are using it wrong. This technology is the most financially accessible technology that has been introduced in years. Comparatively, if you want to put in an ERP system, it is going to cost you a lot of money."
Compared to the legacy technology costs that drinks businesses have absorbed for decades – ERPs, warehouse management systems, custom integrations – a £20 monthly subscription to a professional AI tool is essentially a rounding error. The barrier is not cost. It is understanding.
What AI can actually do for a drinks business
Theory is one thing. The session was most alive when it got specific. What does AI actually do, today, for businesses in this industry?
Customer communication and sales admin. Laura Rosenberger, operational AI consultant and founder of Superposition, described a pattern she sees repeatedly: salespeople pinging emails into a shared inbox, where a team manually looks at each one and types information into systems. "An agent can do all of that stuff. The agent can communicate back with the customer or the sales team. That takes the most time-consuming thing away from salespeople so they can go and sell."
Data quality and reconciliation. Nick Martin highlighted something that will resonate with anyone who has ever tried to clean a customer database in a wine business. "AI will spot issues in data coming out of background systems — duplicate customer codes, broken links between orders. Being able to spot those things is just transformational."
Meeting notes, follow-ups, and institutional memory. Paul Cartwright described how his team transcribes all meetings automatically, creating a searchable record of every conversation — genuine institutional memory that any team member can query.
Onboarding and back-office process automation. James Grant offered a striking personal example. "I built a demo for a client last week which automated entirely a 30-stage new joiner onboarding process. Took me 20 minutes to do that in Claude's co-work. Teeing up emails, writing contracts, sticking things in diaries. And I am the least technical person in my team."
Scaling without the need to hire as many people Fero's own clients offer perhaps the most striking numbers. One Irish client, in the past 12 months, since adopting AI-powered tools built by Fero, grew their top line by 572% and their profit by 410% with the same team. No new hires required.
Website and customer discovery. Nick Martin raised the looming question of Google's Universal Commerce Protocol (UCP). UCP is AI-powered shopping that allows customers to describe what they want and have an agent find, select and purchase it on their behalf. For wine businesses without rich, structured, accessible data, this is a threat. For those who invest now, it is an opportunity.
The automation vs. AI distinction
One of the more nuanced points of the discussion came from Laura Rosenberger, who made a distinction many businesses overlook: a lot of what companies call AI transformation is actually just automation. She points out the reality is automation could have happened years ago.
"Probably two-thirds of what we find in businesses in terms of the inefficiencies they've got at the moment could have been solved by automation five years ago. AI is now just sexy. Boards are speaking about it, they bring us in, and we have to say — sorry to disappoint." — Laura Rosenberger, AI consultant
This is not a reason for cynicism. It is a reason for honesty. Before any business rushes to deploy large language models, it is worth asking: which parts of our operation are still running on manual processes that even basic software could handle? In most drinks businesses, the answer is sobering.
That said, what AI enables goes significantly beyond what traditional automation could achieve. This is particularly true in its newer 'agentic' forms, where models are used to actively do things rather than just produce content. Rosenberger noted that the amount a non-technical person can now automate themselves, without a software developer, without a large IT budget, is 'out of this world.'
Who should own this in your business?
The session was asked a question that will ring true for every managing director of a mid-sized wine merchant or regional distributor: who should be driving the AI agenda? Most of these businesses don't have a CTO. They barely have an IT function.
The answer from the panel was consistent, if perhaps surprising: not the IT department.
"This should not be led by IT," said Cartwright. "This is a business problem, not an IT problem."
Rosenberger offered the most structured answer — a two-pronged approach: a smart, operationally-minded person within the business who can become the internal AI champion, combined with a leadership team that is genuinely committed to transformation.
"It’s really important that you have a leadership team that is fully behind it (adopting AI). Because what you are asking of your company is transformation. They need to be inspired. The need to feel comfortable. The companies that have started making good progress have had a founder or a CEO who's had that 'aha' moment where they've seen the light and they've got that vision for what it means for the company." — Laura Rosenberger, Superposition
The internal AI champion, she stressed, does not need to be technical. Operational experience may be more valuable than a coding background. "When I looked at how AI engineers were approaching building AI — identifying the right problem, breaking down the right process, getting the team on board — I thought: that is precisely ops."
Does everyone need to become an AI expert?
The panel was divided on whether every person in a drinks business needs to develop AI literacy.
Fowler pushed back on the power-user expectation: "I actually argue no. If your talent is selling wine to people, having relationships — those people should be able to use AI tools to help with that, but those tools should come to them. They shouldn't have to work out how to optimise a prompt."
Martin drew a useful analogy: "Think about how good you need to be at Excel. You've got people within your businesses who are total Excel wizards. Then you've got people who just need to know how to read them and update them. You're going to have two types of people with AI as well. But in the same way that you wouldn't employ someone who said 'I just don't use Excel', that same thing is going to happen with AI."
The consensus was that everyone in a business needs a working literacy of AI. Not everyone will need to be an expert, but everyone will need to engage. The person who says 'AI just isn't part of how I work' will, within a few years, be in the same position as the person who said that about the internet.
The stakes are higher
There is a temptation, when surveying a room full of people who love wine and love the drinks trade, to conclude that the human elements of this industry are so irreducible, so resistant to automation, that perhaps it doesn't matter if AI adoption is slow. Relationships are relationships. Terroir is terroir. A great palate cannot be trained on data.
This is true, but it is also beside the point.
The relationships and product knowledge are the things that make the drinks industry worth preserving. But they are currently sitting on top of operational infrastructure that is, in many cases, decades old. The back-office systems, the manual processes, the data silos, the disconnected teams are not part of what makes the drinks trade special. They are the cost of doing business in an industry that has been slow to digitise. And that cost is now being measured not just in inefficiency, but in competitive disadvantage.
"It's not like walking into Tesco and pulling something called AI off the shelf. It's about partnering with people that are using these capabilities to make their products more efficient." — Mitch Fowler, Fero
The drinks industry has seen disruption before. Each time, there were businesses that adapted and businesses that didn't. The ones that didn't generally had good reasons. It may have been tradition, a relationship, a sense that some things are too important to change. They were right that those things were important. They were wrong that they were reasons to stand still.
The next two years
The session closed with a forward-looking question: what does the next two years look like?
Fowler pointed to a shift already underway in how people interact with AI tools moving from typing to speaking, from chatting to commanding. "There'll be a modal change in terms of how we interact and how it's meant to act. The way developers are interacting, they're not actually using their keyboard anymore."
Martin raised the prospect of AI-enabled changes to how wine is actually discovered and purchased, like the use of Giggle’s UPC. For businesses without rich, structured, accessible data about their products, this is a threat. For those who invest now in getting their data in order, it is a significant opportunity.
And beneath all of this, there is a more fundamental shift. Fero's entire engineering team now generates around 90% of its code using AI. "No one's writing any code at Fero anymore. Three years ago, that was not the case." What this means is not that programmers have been replaced. It means that the people Fero employs now spend virtually all of their time understanding customer problems, designing solutions and making decisions. They are thinking more, not just doing. The routine has been stripped away. What remains is the work only humans can do.
That is the promise AI holds for the drinks industry, too. Strip away the admin. Automate the reconciliation. Let the agent handle the inbox. Free up the buyer to buy. Free up the salesperson to sell. Free up the merchant to know their customer.
A starting point
For anyone leaving this article thinking 'this sounds important but I don't know where to begin,' the panel offered practical starting points that require no technical background, no large budget, and no IT department.
Sign up for Claude Pro, or ChatGPT Plus, or Microsoft Copilot — whichever you prefer. It costs roughly the same as a bottle of decent wine per month.
Then, as Cartwright suggested: just ask it a question. Not about wine. About your business. Let it take you down a rabbit hole. The point is not to immediately automate everything, it is to begin understanding what these tools can and cannot do.
The drinks industry's greatest strength is also, at this moment, its greatest vulnerability: it is built on trust, on relationships, on human expertise that no algorithm can replicate. But trust requires showing up. Relationships require time. Expertise requires focus. AI cannot give you any of those things directly. What it can do, is give you back the time, the attention and the energy that administrative drag has been stealing from your business for years.
