As written by ChatGPT, based on the input of an interview with Benjamin Bado, Data and Innovation Manager at Puratos.

Meet Benjamin Bado, Data and Digital Innovation Manager at Puratos. Such an innovative example is the AI Chatbot we built to help their clients. For those unaware, Puratos is an international company that specializes in the production of ingredients for the bakery, patisserie, and chocolate sectors with a Belgian twist. They’ve been providing the culinary world with various services and ingredients for over a century across 100 countries, including Belgium. For a taste of their offer, hit up their website. But for now, we’ll focus on the Proof of Value we built to improve efficiency in the way their clients find what they’re looking for.

Data-driven dilemmas

Picture this: a vast array of over a hundred types of complex (food) products, of which even some of the most experienced bakers don’t know the exact product name. This wide offer of products often makes it difficult for customers to find what they’re looking for. So, enter MbarQ. Fueled with innovation and a strategy to tackle this challenge, we came up with a chatbot that’s not just a virtual assistant, but a culinary colleague as well.

“We have more than 100 different types of products, of which some are hard to find by the customers. So if someone doesn’t know the right name, they can search the product based on an issue or description.”

[ Benjamin Bado, Data and Digital Innovation Manager, Puratos ]

Developing a digital sous-chef

The goal? Simplify the quest for the right ingredient for whatever you’re baking. If a customer can’t recall the name of the exact flavor they need, no worries. Ask the chatbot a question or describe the product, and the chatbot unveils a wide range of products with specific characteristics and application tips. It’s like having a personal sous-chef and store worker simultaneously at your fingertips.

Refining the recipe

In a blink-and-you-miss-it timeframe, we conjured the model. But, as every culinary adventure has its hurdles, we had to hit pause due to a data quality hiccup. No sweat, though. It was all about refining the recipe for success.

So that’s what we did with the help of Puratos’ Product Manager. He interrogated the Language Learning Model to see whether it works accordingly as a Proof of Value. Armed with valuable insights, we were able to see what the challenges were and rearrange and improve our data. With the right result, because fast-forward to now; Puratos is gearing up for a grand website debut of our Chatbot in Q1 2024.

The MbarQ flavor

So, why MbarQ? Puratos holds a close partnership with Microsoft, who recommended us as the Go-To Partner for Generative AI. Initially, they recommended Puratos to work with ChatGPT, which created a dilemma: deep dive in a new technology or trust a partner with profound expertise. And that’s what they did.

In our turn, we gratefully took in the offer with both hands by proactively building a use case specifically for them, working on their needs. And that did the trick, as we could deploy faster than when they had to explore the advantages of AI internally. Besides being fast and result-driven, we had a good collaboration with their internal team of Data Scientists and the Product Manager, making the entire operation run smoothly.

“We chose a partner that is already experienced with the tech. (…) MbarQ was really proactive and already built a use case specifically for us. So they worked on our needs and we’re going to continue working with MbarQ.”

[ Benjamin Bado, Data and Digital Innovation Manager Puratos ]

Results

So, what did we take away from this project? To start off, we offered Puratos a crash course in GPT tech: knowing its quirks and understanding its costs. Or in other words, exactly what they didn’t want to find out themselves, because it would take them too far.

Secondly, we helped them arrange their data. Puratos always knew their data was not flawless. During the project we explored in what way we needed to adjust their current data to make sure GPT could use it effectively.

And finally, we built the Proof Of Value of the Chatbot. After testing it extensively, we could work on quite a few ‘lessons learned’. This added to further understanding the technology and explore its immense potential. Although Puratos invoked our expertise, it’s always important to test or try what you’re doing at some point, which is the next step in this project. So that’s what we did. The next step is to build an MVP, which we’ll be testing on Puratos’ website, starting from Quarter 1.

Conclusion: To chatGPT or not to chatGPT?

To conclude, let’s take a look at why you should use ChatGPT and whether it’s yours to use. First of all, it’s the Usain Bolt of adaptation, bringing innovation to your doorstep. Meaning, it’s really fast accessible for everyone who needs to work with it, which was important to Puratos as well. Early access wasn’t a choice; it was a necessity. They’re always on the lookout for tech that has the ability to change the game. That’s exactly what we did, using Generative AI: Proof of value? Check. Results? Check. Innovative tech that’s helping their business? Double-check.

Secondly, this tech is ready to roll, matured and polished. So don’t wait five years; it’s already the new normal. People use it in their daily lives, and your company should too. That’s what we offered Puratos too.

So don’t miss the feast: the table is set, and the future is served.

Cases