A..I. is coming to the rescue of gastronomy. A young startup has just launched Con.verse, a conversation tool dedicated to enhancing conversation while in gastronomical restaurants.
Aficionados of gastronomy are delighted as they’ve been incessantly alerting recently of the coming demise of gastronomy. Indeed, although cuisine is thriving with social media and visual content, the cultural and conversational aspect counterpart of cuisine is considerably weakening, they argue. « How can you enjoy an excellent meal for three hours if you don’t have conversation? » explains Sara, a long-time believer in combining matter with spirit. « This virtuous circle of feeding thought is in peril » she concludes.
The founder of FFT spotted the opportunity whilst dinning in a three star restaurant in California. « After selling my company, I decided to focus on enjoying life, and namely good food, announces Alexandre, still in his late twenties. It was during one of these fantastic meals I realized something was missing. With my girlfriend we spent three very long hours with barely any conversation. We just didn’t know what to talk about. Needless to say, even though the Cuisine was great, overall it was a terrible experience! »
« Everyone has witnessed one day or another a couple at the table screening their phones. Well, it was my turn, he adds with a sigh. »
This is when Alexandre saw an incredible opportunity for conversational A.I.. His idea was to devise a salter sized conversational agent, placed on the table top of nice restaurants, that could help stimulate conversation.
« It has a focus on gastronomy which means the learning curve is faster. We focus on generating conversation topics or explaining the ingredients of the different dishes, sometimes the origins, details Alexandre. »
FFT named the discrete and stylish conversational agent GIA, for Gastronomical Intelligent Amphitryon, or host.
GIA can already start preparing the conversation long before dinner time. When making a digital reservation, customers can provide information on the type of ambiance that is required. They can chose between Royal, 9 1/2 (courses), Rabelesian, Rastputin and Parisian style conversation. There is also a Business mode in preparation which would allow to pull down quick economic data and calculations for table top negotiations.
« The objective isn’t to replace the waiter nor the cellar master, on the contrary, insists Alexandre. Serving and bringing expertise in recommendations remains the domain of those who offer a smile and who will feel the table mood. GIA’s role is to focus on the conversation, something the waiter cannot do, for simple lack of time. It’s a form of reverse reinforcement learning whereby the guests will receive polite nudges to improve their conversational style. When choosing Parisian style for instance, anything too serious will be red-flagged by GIA. And certainly not a kick in the leg adds Alexandre with a laugh.»
Matters of discretion also explain why the waiter will need to stay, even though the order could be placed easily with vocal command. Customers may want to have GIA removed.
« But this also pushes people to really think about their added value: what is it they can do that the software can’t. »
The last challenge seem to be acoustics, as imitating the soft and intimate voice of the waiter still seems to be lagging. « We don’t want GIA to be barking at the table, even though we are working on different voice modes. For instance, a loud annoying laugh option will be available, but not recommended. » As if to reassure future customers, Alexandre goes on to specify that restaurant owners will have the option of turning off certain modes.
FFT sees a global market, starting with high-end restaurants. They have gathered data on length of meal and basket value to focus on venues where people seem to be spending little time.
« There is one three star restaurant in Oslo for instance where they have two servings, which to us is antonymous to fine dinning : you can’t rush the guests. The only explanation we see is that people are bored and finish their meals quickly. We need to put in some futility, some light conversation, concludes Alexander. »
FFT will also be looking at monetizing data with respect to appreciations made live through GIA on dishes served, especially with respect to another realm of Culinary applied A.I., food-pairing. Going even further, he thinks that one day GIA could become the reference in high end restaurant rating, simply because this knowledge and appreciations are live and global and can fall under one standard, instead of having numerous subjective evaluators not necessarily aligned in terms of criteria. « This could disturb the entire rating system! »
And when asked about the old fashion customers who will prefer not to have GIA, Alexander reminds us of the history of Gastronomy. « The only thing they want to leave the restaurant with are souvenirs. No doggy bags, no pictures, no recordings. But they should remember that cuisine became what it is because we had the opportunity to record, build upon and improve, with books and now digital technology. Why couldn’t we do this with our conversation? »
In any event, Alexander has a plate full. The dying art of conversation means his order book is growing fast and will soon have an indigestion if he doesn’t scale quickly!
This sequence is part of the Digital Scenario Ventures by Philippe Cartau for Biztronomy. Though imagined by the author on the basis of existing or imaginable technologies, somewhere in the world there is a strong possibility something similar exists or is in development. An important objective of the Digital Scenario Ventures is to illustrate to future entrepreneurs and managers the difference between the idea and its subsequent implementation.
Comments