Start building your own chatbot now >

As a French chatbot startup in Station F, it was only a matter of time until we started talking about wine. We met with the CEO and co-founder of Matcha, a wine technology service, to understand their product and discover their vision of winetech. 

Could you present yourself and tell us what you’re doing at Matcha?

“Matcha develops data and wine-advice technologies for the wine industry, such as wine sellers, wholesalers and restaurants around the world, to help them sell better.

The story of Matcha started 3 years ago with Eduardo and Benoit, my 2 partners. Benoit and I were both passionate about wine and partnered up to form a sommelier consultancy for restaurants 5 years ago. As wine experts, our families and friends often asked for advice on wine choices; and during our time as sommeliers, many wine retailers and distributors we worked with were very eager to find a solution that offered their customers the best advice on wine.

Seeing the opportunity, we decided to mix wine expertise and AI to create Matcha in 2016.”

Could you describe the chatbot you’ve built?

“We created a bot that can answer any wine purchase question, just as any good wine advisor would do in a wine cellar shop or in a restaurant. We cover more than 10 use cases, and a lot more are ready to help consumers!

The first thing we did when working on the conception was focusing on natural language processing. It was mandatory for us that the bot could provide the same experience – in terms of interaction and natural conversation – that a human expert would deliver.”

Then, we setup an MVP by creating the intents key to wine research on Recast.AI and connecting our API to their service. Recast.AI’s NLP helped us understand the thousands of ways to ask for ‘a light red wine for aperitif, around 10€”, while allowing us to build a quick and adaptable flow. Our API then allows us to translate the data into wine-compliant information and to answer the request accordingly. “

Do you think the oenology industry is ready for AI?

“Definitely! Wine is one of the most beautiful products you can get. It’s full of history: humans have made it for more than 5000 years! However, it is a complex and intimidating product, because most of the time, you don’t have much information on what’s inside your bottle. People who do not work in the industry do not necessarily understand what the cuvée or the vintage is, and can’t really learn without help. This makes the wine industry a bit inaccessible.

It is therefore a challenge to mix such an emotional product with tech and AI, but the result is worth it.

The winetech movement in France is getting stronger. If you ask me, I’d say it focuses on two main things:

– help wine makers produce better wine, by challenging the vinification and wine making processes, or assist winegrowers with issues in their vineyards

– simplifying and boosting wine distribution, which is what we do at Matcha.”

How is the product being received by oenologists and wine merchants?

“Well! Our product market fit is good. We launched the product 3 months ago, after extensive discussion with our network of distributors about their needs and expectations. Things are going good! The first metrics show a doubling of conversion rates.

We have come to a high-value niche market where the competition is emerging just like us. That means we are not a threat to anyone, have a lot of space to cover, and a lot of freedom!”

Why a chatbot?

“Buying wine should always be a pleasure, centered around quality, personalised and interactive advice. To offer this kind of service digitally, nothing comes closer than a bot. A chatbot offers an experience focused on natural and efficient conversation, and that is exactly what we wanted to offer.”

You can try out Magnum by Matcha on Messenger or visit their website. Many thanks to the team for their valuable insights! Share your thoughts in the comment section. 


Also published on Medium.

Want to build your own conversational bot? Get started with Recast.AI !

  • universy

    Thanks for breaking it down like that!

This site uses Akismet to reduce spam. Learn how your comment data is processed.

This website uses cookies

Please confirm that you accept cookies to monitor the performance of our website