Mondra AI spurs food industry net zero as ChatGPT power endures in Microsoft

Key Takeaways

Mondra has launched Sherpa, an AI-powered assistant designed to help the food industry achieve net zero emissions by automating Life Cycle Assessment (LCA) and analyzing environmental impacts of products using Microsoft Azure's secure servers.

Sherpa enables food retailers to make data-driven decisions by identifying high-impact areas, optimizing resource usage, and running 'what-if' scenarios to explore the effects of ingredient changes on overall environmental impact.

Mondra's platform addresses the significant challenge of measuring and managing Scope 3 emissions in the supply chain, providing insights that help businesses collaborate with suppliers for performance improvement and competitive advantage.

Product footprinting and supply chain decarbonization platform, Mondra, has launched Sherpa, a transformative AI-powered assistant that comes as part of the company’s imperative in bringing the food industry towards a net zero food system. The chatbot comes built using Microsoft Azure OpenAI’s secure servers, in a timely reminder of Microsoft’s relationship with OpenAI.

With Microsoft recently unveiling new AI agents, the focus on the vendor in recent weeks has been more on its Copilot offering rather than anything ChatGPT. With tech media highlighting a seeming messaging change when it comes to the Microsoft-OpenAI relationship, there’s still no change to the basis of Azure OpenAI Service, which provides REST API access to OpenAI language models, resulting in tools such as Mondra’s Sherpa assistant.

Operating on secure Microsoft Azure servers and integrated into the Mondra platform, Sherpa leverages the “best in class” GPT-4o model to augment Mondra’s existing capabilities, automating Life Cycle Assessment (LCA) at scale, previously an “impossible” task, and allowing food retailers to comprehend the environmental impacts of every product they sell. 

Through a chatbot interface, Sherpa allows users to query and analyze their own data in the Mondra platform, and can pinpoint high-impact areas, optimize resource utilization, and model potential outcomes.  For example, Sherpa lists products with the best environmental performance or identifies ingredients with the highest emissions. It can also run “what-if” scenarios to explore how changes in ingredient utilization might affect their overall environmental impact, empowering businesses to make data-driven decisions, reducing their environmental impact and improving operational efficiency.

Speaking with ERP Today at Microsoft AI Tour London from Mondra were Jason Barrett, founder & CEO, and Marco De Sanctis, CTO. The execs explained how the biggest challenge food companies face in getting to net zero is measuring and managing Scope 3 emissions (those that occur in the supply chain).

With the newly-launched Sherpa, they believe category teams can use its insights to collaborate with suppliers on performance improvement, accelerating reduction of Scope 3 emissions, reducing business risk in an ever-regulated market, and enabling competitive advantage. 

De Sanctis, crowned by Redmond as a Microsoft Most Valuable Professional for the last 13 years in a row, explained the capabilities of Microsoft’s Azure cloud platform allows it to keep not just cloud compute costs down but also the carbon footprint of its own computing infrastructure.

“Microsoft Azure provides us the capabilities of [measuring] our own carbon footprint,” he adds, with Mondra keen to uphold the net zero mission also for itself, not just its customers.

On the OpenAI front within Azure, De Sanctis underlines the impact of its generative AI (GenAI). For Mondra, more traditional AI and ML technologies of the neural network sort help its platform categorize hundreds of different categories for food ingredients. This is how Mondra currently monitors more than 30k own label products and over one million unique ingredients.

With GenAI, a “completely different level” has been achieved, for example with translating technical texts from English. He also gives the example of how semantic searches can generate answers on what products are suitable for vegans.

“You cannot answer that question by looking at the ingredients as you don’t have a reliable flag saying Vegan/Not Vegan,” the CTO explains.

The proof, though, as they say, is in the pudding, and Jason Barrett reveals that an own label ready meal produced by a leading UK grocery retailer was reformulated using Mondra, and just one ingredient change in the lasagna reduced the carbon footprint of its production by 18%.

“And because they sell so many this is the equivalent to powering 2000 homes for a year,” he says.

“Our platform can analyze the LCA data of 30k own label products in sub 4 hours daily,” the CEO continues, “a process that would have previously taken over 100 years – and we know this because a leading UK grocery retailer tried it.”

Food for thought, indeed. Read more about Mondra and its smart Sherpa here.