Biodiversity blind spots: Why the next frontier of sustainability needs GeoAI
For years, biodiversity sat just outside the core of most corporate sustainability strategies. It was acknowledged but rarely measured, recognized...
Sustainability compliance used to revolve around documentation. Was a certificate available? Could a supplier confirm something in writing? Was an audit conducted, even if infrequently?
That model is no longer enough.
Across industries, the pressure to transition from reported intentions to verified impacts is intensifying. Regulations such as the EU Deforestation Regulation (EUDR) are raising the standards for traceability. Simultaneously, investors, consumers, and civil society are demanding higher levels of proof and posing tougher questions about how sustainability claims are formulated.
However, for many sustainability teams, the underlying systems have not kept up. Data is still gathered manually. Field audits are expensive and time-restricted. Supply chain visibility is scattered across various regions, crops, and compliance requirements. Despite the abundance of geospatial data, a significant portion remains underutilized or inaccessible without specialized tools.
In a recent webinar hosted by Innovation Forum, panelists from Google, Unilever, and Picterra explored how technologies like Geospatial AI (GeoAI) and remote sensing are helping companies tackle these challenges. What emerged was not a story of simple solutions but one of hard-earned insight. The discussion focused not only on what’s technically possible but on what it takes to build systems that are credible, scalable, and prepared for the demands ahead.
A recap from Innovation Forum is also available to read here.
One of the most consistent themes in the discussion was the disproportionate amount of time and money spent trying to collect sustainability data, often with limited returns. As Pierrick Poulenas, CEO of Picterra, noted during the session, “Today, 90 percent of the effort is still going into finding the data rather than acting on it.” This imbalance serves as a warning sign. It indicates that teams are stuck chasing fragmented inputs instead of using information to drive decisions.
Part of the issue is that many data systems were built for a different era. Audits remain mostly manual, and declarations are often unverifiable. Even where remote sensing is employed, it tends to exist in technical silos, far removed from daily workflows. Sustainability leaders may be aware that high-resolution imagery is available, but without automation and context, it doesn’t move the needle quickly enough.
This is where Geospatial AI begins to shift the paradigm. By combining remote imagery with machine learning, GeoAI can detect patterns in land use, identify changes, and flag compliance risks with much greater speed and consistency. It transforms vast amounts of data into visual, verifiable outputs that teams can interpret and act upon.
But the panel clarified that the goal isn’t to gather more pixels. The objective is to transform those pixels into intelligence—intelligence that connects to the questions compliance teams are genuinely trying to answer.
While GeoAI provides a powerful tool for monitoring landscapes and identifying risks, the panel unanimously agreed on one point: technology is only part of the equation. Sustainability does not occur in isolation. It takes place in locations, in communities, and within supply chains where human knowledge and context are vital.
Pierrick Poulenas emphasized this repeatedly. “You need people on the ground who know the land,” he said. Whether it’s an agronomist examining a parcel in Brazil or a local NGO monitoring forest boundaries in Southeast Asia, that knowledge adds context to the data. It confirms, calibrates, and builds trust in what the models are seeing from above.
This concept—often referred to as a human-in-the-loop model—ensures that AI systems are not operating blindly. Instead, they are bolstered by expert judgment and informed by local realities. That feedback loop is particularly crucial in sustainability, where conditions change rapidly, and a model trained in one region may not directly translate to another.
The real value comes from integration. GeoAI should support the people already doing the work, not replace them. It can automate the repetitive tasks, flag the unusual occurrences, and scale insights across regions, but its impact depends on how well it connects to the ground truth that only people can provide.
For technology to effectively support compliance at scale, it must be both accessible and aligned with genuine business needs. This was a key message from Alicia Sullivan, Product Manager for Earth Engine Sustainability Solutions at Google, who described how their tools are evolving to meet the expectations of both scientists and sustainability practitioners.
“Earth Engine has traditionally been used in academia,” she said. “We’ve been working on making it easier to use for people who aren’t geospatial experts.” That shift—from high-skill users to operational teams—makes geospatial intelligence relevant across industries. Sullivan also emphasized the importance of partnerships with organizations like the Forest Data Partnership and the Food and Agriculture Organization, where tools are being designed to support smallholders and global supply chains alike.
For Andrew Wilcox, Associate Director of Sustainability, Procurement Strategy and Insights at Unilever, the lesson has been that technology only works when it is integrated into real processes. “There’s a danger of becoming too focused on the map or the platform,” he explained. “The data has to fit into how the business actually operates.” He described how Unilever has utilized geospatial monitoring to support its efforts toward 95 percent deforestation-free sourcing, but noted that progress was only possible because the company connected data to systems of assurance and engagement throughout the supply chain.
Both speakers emphasized the necessity of integration. Whether through cloud-native platforms or open data initiatives, the aim is to make sustainability intelligence more accessible, actionable, and impossible to overlook.
As sustainability regulations evolve, demands on corporate data systems are increasing. Compliance frameworks like the EUDR require more than just risk assessments or supplier statements. They call for traceable evidence, often down to the plot level. Yet, as Pierrick Poulenas noted, the reality is that many companies still rely on inference and estimation to meet those requirements.
“There’s a big gap between what we say we can do with AI and what’s actually operational,” he cautioned. While it may be possible to predict land use with 90 percent accuracy using satellite data and models, that level of certainty is often insufficient when regulations require precise, auditable proof.
Instead of making unrealistic promises, the panel highlighted the importance of leveraging GeoAI to enhance decision-making in more practical and impactful ways. This involves identifying high-risk areas that require deeper verification, facilitating earlier intervention, and refining how internal teams allocate time and resources.
New approaches are also emerging. From foundation models trained on large-scale environmental data to lighter tools that enable users to create custom detectors with just a few examples, the technology is moving toward greater usability and adaptability. But as Alicia Sullivan stressed, technology alone won’t solve the credibility gap. “We need to invest in shared infrastructure,” she said. “Not just better models, but better systems for how data is shared, validated, and used.”
The Forest Data Partnership serves as an example of a collaborative effort to make geospatial data more accessible and trustworthy, especially for nature-based solutions. This highlights that in complex systems, the ability to scale relies not only on automation but also on cooperation.
The shift from ambition to accountability is already happening. Whether driven by regulation, investor scrutiny, or internal conviction, organizations are being asked to do more than just set sustainability goals. They are being asked to demonstrate them.
Geospatial AI, remote sensing, and cloud-based tools provide powerful means to bridge the gap between compliance requirements and operational capacity. However, as the panelists from Google, Unilever, and Picterra emphasized, credibility relies on more than just sophisticated algorithms. It necessitates integration, shared infrastructure, and a commitment to constructing systems that embody the complexity of real supply chains and ecosystems.
Compliance is no longer merely about reporting. It involves the ability to observe what’s happening—and to present it clearly to others. This visibility forms the foundation of trust and is the initial step toward systems that not only track sustainability but also genuinely support it.
For years, biodiversity sat just outside the core of most corporate sustainability strategies. It was acknowledged but rarely measured, recognized...
Carbon markets were created to make climate action more scalable. By assigning financial value to emissions reductions and carbon storage,...
Regenerative agriculture is no longer just a fringe idea. Once regarded as a niche approach championed by a handful of...
For many sustainability leaders, compliance has long been the primary focus: tick the box, pass the audit, and move on.But...
