The Sommelier & The Lawyer: What AI Can’t Replace
When three supermarkets in the French Alps installed A.I.-powered wine merchant kiosks to help customers navigate their large wine departments several weeks ago, it put new pressure on a slice of the wine industry that was already feeling squeezed. It prompted the wine community to ponder a familiar-sounding question: is A.I. trying to replace the sommelier?
The A.I. wine “advisor” at issue, named “Elio,” was developed by retail tech company Etinsy. It is touted as the “first A.I.-powered humanoid advisor dedicated to the wine and spirits aisles in supermarkets.” Designed to respond to casual vocal prompts, Etinsy says that Elio “interacts naturally, verbally.” It mimics how a shopper would interact “with a professional in a specialized [wine] shop.” Johannes Macron, president of the French Wine Merchants Union, lamented in an interview with wine-searcher.com that “large retailers have just realized that [customer] advice is important…Instead of making it human, they’re virtualizing it.”
The Same Questions Facing the Legal Profession
Like sommeliers, lawyers have had to navigate a rapidly shifting professional landscape in recent years. A.I.-powered products and platforms have exponentially sped up data processing. This has emboldened consumers and increased society’s appetite for efficiency along the way. The A.I. revolution has placed extraordinary amounts of information at everyone’s fingertips, leaving the workers with expertise feeling the most targeted and most vulnerable. If a kiosk can identify a great Cabernet that’s under $20 and pairs well with the ribeye already in a shopper’s cart. If a laptop can tell a homeowner whether his plans to build a new garage comply with applicable zoning codes. How important is expertise anyway?
Information Is No Longer the Competitive Advantage
AI Excels at Processing Knowledge
Society had long regarded human experts as the primary sources of knowledge in their respective fields. As A.I. platforms have continued to absorb more data points into their ever-expanding models and as A.I. developers devote more time and resources into producing “agentic” A.I. products and services, focus has shifted more deeply toward contextualizing (and not merely retrieving) information. In this environment, human knowledge itself feels decreasingly relevant each passing day. Soon, A.I. will process all known information and perform all human functions.
Or will it?
Just as A.I.’s rise seems to render some human traits less relevant, it has also highlighted those essential functions that will benefit from more humanity – not less – in our A.I.-assisted future.
While A.I. excels at processing the known and forecasting the expected, much in life is not static but instead evolves to accommodate society’s changing tastes and sensibilities. This evolution is a distinctly human endeavor – and one that is significantly enriched by those with a relevant expertise.
Why Sommeliers Still Matter
Making Wine More Accessible Through AI
One reason Elio is so attractive to supermarkets is that it was designed to recognize and communicate in natural, everyday human vernacular (i.e., “what’s a good red to pair with a steak for under $20?”). Another is that it produces hyper-localized results. Elio monitors the actual in-store inventory at each imbedded location and is designed to factor in a shopper’s meal choice and budget to then suggest a wine that the shopper can actually leave the store with that day.
As the wine industry confronts signs of overall decline (global wine consumption decreased by 3% in 2025 and dropped 4% in the U.S., which is the world’s largest market, over that same span), sellers are generally embracing anything that makes customers more likely to purchase a bottle.
In regions where the wine marketplace has shifted away from small specialty shops in favor of larger stores with more vast on-site inventories, wine producers have worried that selecting a wine can feel overwhelming to the average shopper, who may feel safer skipping a wine purchase altogether rather than arbitrarily settling on a bottle that does not wind up jiving with their palate. For some, kiosks like Elio remove the anxiety and make buying a bottle feel like an informed decision and not mere guesswork. The supermarkets that have embedded Elio into their wine departments believe that Elio’s “advice” will lead to more confident shoppers who are more likely to become regular wine buyers.
The Rise of AI Wine Advisors
Elio is the most recent A.I.-based wine innovation, and one that targets the direct point of retail sale – but wine insiders have been trying to harness A.I. to help make wine feel more accessible to a wider community for some time now. Focusing on more than just bottle availability, wine A.I. developers have been inputting industry specific knowledge and data about wine composition and other technical factors into product databases for years. Platforms such as Vivino and Delectable perhaps started the trend as they permitted users to upload pictures of bottles onto their sites which would unlock information such as average ratings and drinker reviews. Platforms since have increasingly refined their designs to advise potential consumers on topics such as food pairings, appropriate glassware, preferred seasonality and more.
The various competing platforms have each been compiling their own proprietary “wine ontologies” over the years, and their developers continue to identify and collect more specific and more nuanced data points designed to produce more varied recommendations. Data about wine characteristics (grape variety, region, flavor profile, aging potential, etc.), food types (protein type, taste characteristics, chemical composition, etc.) and pairing rules (flavor balance, acidity, texture, etc.) are organized and structured by each platform’s own “knowledge graphs” and the categorized input is then interconnected/contextualized to result in “recommendations” far more refined than one could have obtained by consulting the rudimentary physical charts available at some establishments in the old days (red with steak, white with fish, etc.).
Wine-tracking app Invintory markets its own A.I. sommelier, “Vincent,” which boasts “continuous learning” from real-time data provided by each user. Invintory says Vincent’s output becomes more “personalized” over time with increased use and data input. Invintory also claims that Vincent can help its users manage their personal wine collections by “tracking [items such as] aging potential, [rate of] consumption and optimal drinking windows.” It even claims it can factor in “occasion” and “setting” when recommending a wine, taking service specialization to a whole new level.
AI Is Changing the Customer Experience
Wine advice is not merely the realm of savvy consumers who download wine-specific apps, either. Anyone with a cell phone can grab a quick recommendation right from their seat at the bartop or restaurant table via short consult with any generalized A.I. app such as ChatGPT or Google Gemini. Diners can input a restaurant’s wine list into ChatGPT, for example, and ask it to identify the wine most like the one they regularly drink at home. Others may use it to answer basic questions about a restaurant’s available wines before even being asked to place their order.
Where does all this leave the sommeliers? On their toes. As diners arrive armed with more information than ever before, along with greater confidence in their own selection abilities, sommeliers must prepare for more intricate interactions and more detailed discussions with their customers right from the jump. Some restaurants are even requiring their sommeliers to attend classes and demonstrate proficiency with the A.I. tools in the marketplace, particularly the ones that have proven most popular amongst their known patrons. Sommeliers now regularly arrive at tables expecting to find knowledgeable patrons with strong opinions of their own.
What AI Still Can’t Do
So why not just put a QR code on each table and let patrons self-sommelier at this point?
Well, a QR code cannot decant a bottle of Syrah, let alone actually taste the results, sommeliers would argue. Neither can a QR code “read a room” to detect a dining group’s mood – or use intuition to suggest an outside-the-box bottle that turns into a surprise hit.
Human Expertise Shapes the Future of Wine
Wine is dynamic. Even the same wine brand from the same winery experiences shifts from year to year depending on everything from weather conditions to soil quality to changing trends that alter consumer palates and expectations. Sommeliers serve as more than just service staff specialists in this complex ecosystem. They embody the industry’s expertise in a way that helps recognize changes in tastes and directs the emergence of new trends. They can detect changes in the quality or character of a wine that have resulted from environmental, resourcing, or even one-off factors (like an odd grape crop or supply chain blip, etc.). More than simply cataloging the known and recommending the expected, they inject originality and creativity into identifying and selecting wines that helps take wine in unexpected directions. In short, sommeliers’ human contributions are vital to wine’s continued evolution.
For their part, A.I. wine developers acknowledge that the human touch is essential to wine’s evolution and vitality. Invintory’s site notes that “every pairing should be both data-driven and deeply personal – aligning with the consumer’s evolving tastes” while also concluding that “the best outcomes happen when technology and personal preference work together.”
A.I. can deepen the conversation about wine, widen the data set used to advance the quality and character of wine, and even increase the public’s confidence in selecting wines to drink – but the trends that keep wine current cannot emerge from hard data alone.
Sommeliers are wine society’s most trusted representatives for a reason. They absorb the hard data, filter it through lived experience, and then calibrate their recommendations to accommodate their patrons’ – and the overall consuming community’s – tastes and preferences. Rather than dying off in this new A.I. era, the human elements of trust, integration and taste may be more alive than ever.
The Legal Profession’s AI Reckoning
Perhaps lawyers and sommeliers should exchange notes.
From the moment this new A.I. boom gained traction, law firms have been asking existential questions about their business models and even law practice itself. Back when Westlaw and LexisNexis revolutionized legal research with electronic and then natural-language case searches, lawyers largely marveled at the increased speed with which they could now identify more tailored citations and drop more up-to-date caselaw into their legal briefs. This increased access to information mostly made lawyers feel empowered, not diminished.
The A.I. wave of today – brought on by ChatGPT, Claude, Harvey and continual newcomers – has hit much differently in the legal community. Today’s A.I. agents are designed to be conversational and increasingly contextual. Armed with vast libraries that contain all statutory law and case precedent known to man, these agents can – and do – produce lawyer-like work product. Today’s A.I. can write an opposition to a lawyer’s motion before the original brief has even been filed. It can stand in for a district judge or an appellate panel and pepper a lawyer with questions in advance of a hearing. It can summarize testimony in real time and can even search records to find inconsistencies in a witness’ statements before he/she has even stepped off the stand. If used ethically and securely, these tools can take lawyer efficiency and preparedness to extraordinary new levels.
Lawyers and firms recognize both the current and potential changes this burgeoning A.I. era represents for law practice. Many have formed committees to vet, test and implement A.I. tools that add value and increase output and most have policies in place to monitor and police the ethical use of these new products.
Clients and potential clients recognize this new landscape, too. Many (most?) have already obtained A.I.’s “legal opinion” (along with a quantum assessment, if relevent) of their case before they even request an initial consult – and lawyers, like sommeliers, have come to expect a more refined discussion as clients arrive with more detailed inquiries upon first engagement.
As lawyers keep pushing to keep pace with A.I.’s progressions, though, many still worry about being trampled under a coming stampede. A.I. legal products seemingly grow more sophisticated by the day as agentic tools are fed more data points and grow more proficient at completing more complex tasks.
If A.I. can conjure up an opposition to a motion for summary judgment, then why not have it write the original brief? And if A.I. can draft both parties’ pleadings, why not just have it settle the dispute? A.I. can already access the same vast libraries of law and precedent as the judges, so why not just appoint A.I. as the permanent judge of everything?
Perhaps things are already headed that way. Several companies are testing and marketing A.I.-exclusive arbitration platforms in industry sectors where mandatory arbitration clauses are commonplace and expected, for example. Rather than paying a legal team and waiting months for a human arbitrator’s decision, these platforms boast that their bots can process the dispute, apply all the known precedent, and pump out a ruling in mere minutes.
“Hallucinations” and fabricated citations have made notable headlines and highlight the obvious need for caution in the current climate. A.I. developers will continue working to address this issue, however, and they will likely succeed in eliminating or vastly reducing these errors someday.
Then what?
Are hallucinations the only drawback to letting A.I. draft all legal briefs? Will lawyers become fully expendable if/when these faulty citations are held in check?
Perhaps the increasingly popularity of Chilean Cinsault, Frappato and other lower-alcohol, chillable reds offers a clue. Sommeliers helped usher these wines into the mainstream as lighter options during the increasingly hotter months of the year. They’ve also overseen a shift from heavily oaked Chardonnays to more textured, “food-friendly” whites as consumers express desire for options with more “personality,” and have aided a rise in English still and sparkling wines that have benefitted from warming growing conditions.
Like wine, the law is never finished. It evolves along with the society it is written to govern. Just as the right wine choice depends on myriad factors beyond the bottle itself – the meal, the occasion, the weather, etc. – adjudication of a legal dispute similarly depends on many factors well beyond the law as it is written down. Each dispute comes with its own unique set of facts. Resolution turns on how the fact-finder views the facts and how the court applies the law to them. They serve as society’s representatives in resolving each dispute, but they are not alone.
What Makes Human Lawyers Irreplaceable
Lawyers do more than look up the law and cite to precedent. They are members of society. They feel the effects of the laws on their own lives and observe the ways that laws impact the lives of their neighbors. They sense changes in society’s overall views and sensibilities, and they integrate their lived experience with their legal expertise to frame the facts and issues for the fact-finder and the court.
Human lawyering is vital to ensuring that the law continually evolves with society’s changing expectations because human lawyers live through, experience and observe these changes themselves. Human lawyers are uniquely qualified to filter the facts through an updated frame of reference and to craft the unique, previously unmade arguments that precedent alone cannot anticipate. Innovating and, at times, bold lawyering has guided the law through changing societal views on everything from racial equality to reproductive rights to gambling on sporting events.
This is a big reason why some law firms and law schools have enacted outright bans on lawyers and students using A.I. to draft legal briefs. A lot like decanting a wine, drafting requires the integration of personal and technical skill to produce a truly exquisite result. Reflection on the human condition requires human contribution.
The Future Is Collaboration, Not Replacement
A.I. may elevate the initial conversation between a client and a lawyer or a customer and a sommelier. It has certainly altered the way these relationships are formed and maintained. But just as an app can never actually taste the Viognier that it recommended for pairing with a crab dressing-stuffed pork chop, it can never feel the effects of the law on people’s lives, nor can it anticipate how society will evolve in response to these lived experiences. This kind of evolution is a job for the living.
And there’s help wanted.
