VentureBeat presents: AI Unleashed – An exclusive executive event for enterprise data leaders. Network and learn with industry peers. Learn More
Whether or not you subscribe to the “butterfly effect” — coined by meteorologist Edward Lorenz in 1961, the theorem states that a butterfly flapping its wings in Brazil could, through a chain of various interlinked meteorological events, ultimately cause a tornado in Texas — there’s no denying that events far from us on the other side of the world have an impact on our lives.
This is especially true in the agricultural supply chain, where companies that sell coffee in my home country, the United States, typically import the beans from far-flung locations with climes better suited for growing them, such as countries in Central America, South America, and Subsaharan Africa.
But consumer packaged good companies (CPGs) and food processors have until now had to rely on a combination of different methods for keeping track of their supply chains, including waiting on direct word from the suppliers themselves, to using software to monitor weather and news events in the countries their products originate from, and keeping track of the supply chain in non-specialized software such as PowerPoint and Excel.
A new startup thinks it has developed a better way: today, Helios, co-founded by a former at Boston Consulting Group (BCG) and a former award-winning machine learning engineer at Google, is introducing Cersi, which it bills as “the world’s first supply chain AI analyst.” Cersi is a conversational AI chatbot, similar to ChatGPT or Claude 2, but specialized for the agricultural supply chain.
An exclusive invite-only evening of insights and networking, designed for senior enterprise executives overseeing data stacks and strategies.
Users at CPGs can feed Cersi their supplier information and location, and it will automatically surf the web and track relevant updates for that location’s weather forecasts, natural disasters, the effects of climate change, and other news events that could impact the supply chain.
“We started it because we wanted to make sure that preventable food supply shortages never happen again,” said Francisco Martin-Rayo, Helios CEO and the co-founder who was a former principal at BCG, in a video call interview with VentureBeat. His co-founder, Eden Canlilar, serves as chief technology officer at Helios, but worked at Google as an AI/ML engineer, where she won the 2021 Emerging Technologist Abie Award.
The company is also announcing it has raised $1.85 million in pre-seed funding from Supply Change Capital with participation from January Ventures.
‘Billions’ of climate and weather data points
Martin-Rayo told VentureBeat that Cersi is part of the broader Helios platform, which works by ingesting “billions of climate, economic, and political signals.”
“Plus, we track all these force majeure and catastrophic events like earthquakes, tsunamis, and floods in realtime to give your global suppliers a realtime, highly-customized view,” he continued. “If you’ve got 150,000 suppliers you import from, we can give you a snapshot in realtime of which ones are at the highest risk of disruption.”
Helios crawls 60,000 news sources every day to help inform its platform and Cersi’s responses, keeping them highly updated and relevant, according to Martin-Rayo.
In a demo view showed to VentureBeat over the video call interview, Martin-Rayo showed off the “supplier dashboard” that Helios’s initial small group of beta customers have had access to since March 2023. It allows customers to input their supplier name, the commodity that is being supplied, and the location from which it originates.
Then, Helios provides a score and shows if the commodity is at low, medium, or high risk of disruption from natural or human-driven causes, as well as an explanation of what that score was assigned.
The product also shows a map of the world that represents a customer’s entire global chain as points on it, which the user can click into to learn more about that particular supplier and the events impacting them.
Where Cersi comes in is as a conversational assistant. Now, instead of a user having to navigate this dashboard all on their own and copy and paste the data over into a report, Cersi can fetch it and prepare it using natural language queries and providing responses in similarly natural language.
“The reason this is so special,” Martin-Rayo asserted, “is that historically…these [CPG or wholesale supply] companies would have these existing contracts with consulting firms, where they would pay tens of thousands to hundreds of thousands for reports on this information, and it might take a couple of days or weeks. Now, you’re able to get it in seconds.”
Innovating new weather analysis AI
Coming from the white-collar worlds of business and software, Martin-Rayo and Canlilar had lots to learn about agriculture when developing Helios — and when building Cersi in particular.
“Frost occurs through a precise combination of humidity and temperature,” Martin-Rayo said. “Liquid inside plants will burst open, and that’s why it’s damaging for agricultural crops. We’ve had to build different proprietary [machine learning] models around growing seasons.”
Martin-Rayo said partnerships with former leaders at Coca Cola, Starbucks, and Aldi have helped inform their product and its data analysis.
Unlike other companies that have chosen to pursue a strategy of springboarding off leading commercial models, acting as “wrappers” of OpenAI’s GPT-3.5 for example, Helios is leveraging upon source and its own internal ML chops.
Asked about the technology underlying Cersi, Martin-Rayo told VentureBeat: “We’re not using GPT. It’s been a combination of a lot of proprietary stuff and open source. It’s been an enormous lift to be able to build it.”
Though Helios has set its sights on improving supply chain resiliency and “making sure that preventable food supply shortages never happen again,” Martin-Rayo also admits that commodities traders who can take advantage supply shortages to make money are also among the users of his company’s technology, but says they too could help ensure a more secure food supply chain.
“I think the mentality from the customers that we talked to is much more along the lines of how can we be much more proactive and work with our suppliers to mitigate some of these risks,” he explained.
Right now, Martin-Rayo said that “a majority” of Helios’s early customers “have been food processors and CPGs.”
VentureBeat’s mission is to be a digital town square for technical decision-makers to gain knowledge about transformative enterprise technology and transact. Discover our Briefings.