The Scene From Above Podcast

In the final episode of Season 16, Julia Wagemann speaks with Phil Cooper, Geospatial Global Tech BD for the Amazon Public Sector team, and Nadine Alameh, Founder of LunateAI and former CEO of the Open Geospatial Consortium.

The conversation closes the GenAI for Geospatial Challenge takeover by looking beyond the individual prototypes and asking what this moment means for the wider geospatial industry.

Phil reflects on the evolution of Earth observation from a time when researchers might spend months analysing a single satellite image to today’s world of petabyte-scale data, cloud infrastructure and AI-enabled analysis. Generative and agentic AI are now helping organisations engage with geospatial data without needing years of specialist experience, opening the field to new industries, users and ideas.

Nadine describes this as a moment in which anyone with a strong idea can begin turning it into a prototype, solution or business. She also reflects on the Challenge’s role in demonstrating the art of the possible, lowering barriers and supporting teams working on meaningful problems across climate, health, oceans, infrastructure, insurance and spatial intelligence.

The episode explores the patterns that emerged across the 14 winning teams. Unlike earlier experiments focused primarily on the technology itself, many of this year’s projects were grounded in specific domain challenges, including methane monitoring, malaria planning, cocoa supply-chain transparency, marine intelligence, disaster risk and insurance.

Phil and Nadine also discuss the rapid move from generative AI towards agentic systems, multimodal geospatial foundation models and natural-language interfaces. As these tools become more operational, the conversation turns to trust, data governance, standards, provenance and the need to help users understand where information comes from and how it has been processed.

The episode also considers which sectors may adopt geospatial AI first, why energy and finance are already moving quickly, and how geospatial intelligence could become embedded into everyday decision-making across industries.

Host: Julia Wagemann
Guests: Phil Cooper and Nadine Alameh
Organisations: AWS and LunateAI
Episode focus: The future of generative AI and geospatial
Series: Scene From Above, Season 16 – GenAI for Geospatial Challenge Takeover

In Episode 3 of Season 16, Sabrina Szeto speaks with Dr Cormac Purcell, astrophysicist turned geospatial scientist at BolgiaTen.

BolgiaTen’s project, developed through the GenAI for Geospatial Challenge, focuses on marine domain awareness: helping users understand where vessels are, what they are doing, and whether their reported movements can be trusted.

The project addresses a major challenge in maritime intelligence. Automatic Identification System, or AIS, data is widely used to track vessels, but it is based on an open trust system and can be affected by GPS jamming, spoofing, or deliberate manipulation. This matters because incorrect or falsified vessel positions can have real consequences, from insurance risk and sanctions compliance to dark port calls, ship-to-ship transfers and activity in protected marine areas.

In this conversation, Cormac explains how BolgiaTen combines AIS data, vessel context, ownership and maintenance information, and Earth observation data from open and commercial sources to build a more reliable picture of vessel behaviour.

The episode also explores the role of generative AI in the project. BolgiaTen uses foundation models to accelerate vessel detection in Earth observation imagery and generative AI to support a natural language interface for the analysis pipeline. What once required specialist analysts to process imagery, interpret results and answer business questions can increasingly be supported through an agentic interface that helps orchestrate the workflow.

Cormac also shares how AWS cloud credits helped the team move beyond an isolated prototype towards a more professional deployment pathway, closer to the environment used by their partner Lloyd’s List Intelligence. The credits supported experimentation, integration and a higher technology readiness level prototype that can be tested with customers more quickly.

The conversation also covers the importance of customer education: helping business users understand what Earth observation can and cannot do, how resolution, cadence, coverage and cost shape what is possible, and why clear communication of uncertainty is just as important as the underlying science.

Host: Sabrina Szeto
Guest: Dr Cormac Purcell
Organisation: BolgiaTen
Project focus: AI-enabled maritime domain awareness
Series: Scene From Above, Season 16 – GenAI for Geospatial Challenge Takeover

In Episode 2 of Season 16, Julia Wagemann speaks with Reda Haidar, Commercial Director at RASID.

RASID’s project, GoPilot, is an agentic Earth observation platform developed through the GenAI for Geospatial Challenge. It allows users to ask questions in natural language and receive clear, evidence-backed geospatial intelligence without needing to manually search satellite archives, download imagery or run complex workflows.

The project began with a focused use case around methane monitoring, using RASID’s Methane Mapper system to detect methane emissions from satellite imagery. During the Challenge, the team realised that the same architecture could support a much wider set of Earth observation workflows, from agriculture and environmental monitoring to infrastructure analysis and change detection.

In this conversation, Reda explains how generative AI acts as the reasoning layer in GoPilot: understanding user intent, selecting the right tools, retrieving relevant data, coordinating analytical workflows and synthesising results into actionable answers.

The episode also explores the role of AWS cloud credits in helping the team experiment at scale, including the generation of more than 100,000 synthetic methane plumes, model training, and the development of an agentic geospatial platform using AWS tools and infrastructure.

Host: Julia Wagemann
Guest: Reda Haidar
Organisation: RASID
Project: GoPilot
Series: Scene From Above, Season 16 – GenAI for Geospatial Challenge Takeover

In the first episode of Season 16, Sabrina Szeto speaks with Nana Akua Afra Owusu-Addo, Technical Writer for the Earth Observation Team at BigData Ghana Limited and Project Lead for ForestTrace AI.

ForestTrace AI is an AI-powered forest and land use intelligence platform developed through the GenAI for Geospatial Challenge. The platform helps governments, environmental agencies, cocoa sector actors and sustainability teams monitor forest change, detect deforestation risk and turn satellite data into clear, decision-ready insights.

In this conversation, Afra shares how the team is using generative AI to make geospatial intelligence more accessible to non-technical users, including policymakers, sustainability officers, field teams and local stakeholders. Rather than simply showing maps and dashboards, ForestTrace AI helps users ask questions in natural language, understand what the data means and identify what action may be needed.

The episode also explores BigData Ghana’s journey through the challenge, the role of AWS cloud credits and technical support, the importance of responsible data governance, and the team’s ambition to scale ForestTrace AI into a broader African land use intelligence platform.

 

Host: Sabrina Szeto

Guest: Nana Akua Afra Owusu-Addo

Organisation: BigData Ghana Limited

Project: ForestTrace AI

In this episode of Scene from Above, Julia Wagemann speaks with Matthias Mohr, independent software developer and one of the key contributors to the STAC (SpatioTemporal Asset Catalog) and STAC API specifications.

 

STAC has become foundational to how Earth observation data is discovered and accessed across cloud platforms. But its origins lie in a fragmented landscape of portals, inconsistent metadata, and incompatible APIs.

Matthias shares how STAC emerged from practical needs within the community and how it evolved into a widely adopted standard for geospatial data discovery.

 

Together, Julia and Matthias unpack:

Why STAC was created and what problem it solved

The difference between static STAC catalogues and STAC APIs

How organisations struggle when adopting STAC internally

The role of extensions and interoperability

Where cloud-native geospatial infrastructure may head next

 

A thoughtful conversation for anyone working with large-scale Earth observation data, from analysts querying data, to engineers publishing catalogues, to decision-makers shaping data infrastructure.

 

Host: Julia Wagemann

Guest: Matthias Mohr