Treefera attended Gartner Supply Chain Symposium/Xpo in Barcelona (18-20 May), with the team delivering both a speaking slot and a working lunch - and across both engaging in meaningful conversations around supply chain resilience in an increasingly volatile and observable world.

We asked our Chief Evangelist and Head of Customer Solutions, Gabrielle Bourret-Sicotte, to walk us through some of what the team presented in Barcelona as well as what she heard back from attendees.

The presentation was titled "Food Security Is a Financial Systems Risk." What did you want the audience to leave thinking?

That they already know the risk exists - they have lived through cocoa's 60-year high, Vietnam coffee's typhoon spike, India's overnight export ban - but that they are still pricing and planning against data that captures these events after they have happened. The core message was that the information gap is no longer a technical problem. It is addressable. The question is whether supply chain and risk teams act on that before the next disruption, or after.

You framed three pressures converging: demand, climate and geopolitics. Why lead with that?

Because each one on its own is manageable. The 9.7 billion population projection by 2050 is a long-run trend. A geopolitical input shock - like the 46% urea price spike triggered by Strait of Hormuz disruption - is recoverable. But when demand pressure, weather volatility and input cost shocks compound at the same moment in the crop calendar, the decision window is extremely short. The people in that room manage the supply chains that feed the world. They need to know what is coming, not what happened.

You used the phrase "the decision vacuum." What does that mean in practice?

A crop passes through distinct biological stages: planting, vegetative growth, flowering, grain fill, harvest. The window in which an intervention - a hedging decision, a sourcing shift, a replanning move - makes a material difference closes during the reproductive stage. By the time official crop estimates arrive, that window is gone. The decision vacuum is the space between when ground truth is physically visible in the field and when it reaches the people who need to act on it. That gap used to be measured in months. It is now closeable to weeks.

In your presentation with Jonathan, you shared key questions that Treefera’s customers are asking, and presented how we are using the platform to not only answer those questions but also allow our customers to act in impactful ways. Can you share some highlights?

In Barcelona, we walked through four deep‑learning models we’ve built on top of our multi-sensor stack – each focused on one high-stakes customer question. The point is simple: our First‑Mile Intelligence closes the “decision vacuum” by surfacing yield and growing-condition signals months before they show up in official data, giving clients time to hedge risk and replan supply chains.

• Sugar in India: "Is India about to ban sugar exports again?"
  • The model provided three months' lead time on sourcing, the difference between buying at 14¢ and scrambling at 20¢+ after the market reacted to the 5–7 month delayed official production data.
• Cocoa in Côte d'Ivoire and Ghana: "Where is cocoa under stress – and how bad?"
  • The model acted as a crop-stress early warning, flagging the production shortfall in early March, ten weeks before the market caught up to the 478,000-tonne global deficit.
• Coffee in Vietnam: "Did the floods actually hit the coffee?"
  • The model provided signal-vs-noise clarity, confirming the minimal overlap between flood extent and coffee-growing areas, which saved clients from panic-buying at the 4.5% Robusta price spike.
• Corn in Kenya: "Will this dry spell destroy the harvest?"
  • The phenology-aware forecast flagged 80–100% crop failure six months ahead by weighting drought risk by the maize growth stage, enabling immediate decisions like planting on irrigated land and strategic five-year sourcing replanning.

You also made the case for not acting - on Vietnam coffee. That seems counterintuitive.

It surprised people. Typhoons hit Vietnam's coffee heartland last November. Robusta spiked 4.5% overnight. Bloomberg, Nasdaq and Al Jazeera all ran it hard. The pressure to do something was enormous. But the satellite overlay of flood extent against the actual coffee-growing areas in Dak Lak showed that the crop itself was largely unaffected. The data did not tell you to act. It told you not to. That confidence saved buyers from locking in at a premium that evaporated within days.

And the lunch - what were the conversations like?

More direct than the stage format, as they always are. The consistent theme: people understand the problem - fragmented, delayed first-mile data - but are still waiting to see it integrated into their existing workflows rather than sitting in a separate analytical layer. That is the gap we are focused on closing.