Nothing Left Hanging: Digital Hangar turns video imagery and documents into decisions. Before the moment passes.

Government agencies are inundated with a deluge of multimodal data—video, audio, images, documents, and reporting. When teams can’t triage and correlate it quickly, critical signals get buried, decision cycles slow down, and mission risk increases. 

The Solution: Digital Hangar—a joint innovation from Accenture Federal Services and Google Public Sector. Leveraging Gemini models and Google’s Agent Development Kit (ADK) on the Gemini Enterprise Agent Platform, Digital Hangar uses agentic workflows to instantly process massive data sets, extract structured signals, and connect the dots. 

The virtual hangar 

In the wake of major crises or ahead of high-stakes operations, investigators historically used physical airport hangars to spread out evidence and connect the dots. Today, that dilemma exists in the virtual space. 

What began as a proof-of-concept to assess how early Gemini vision models processed video has now evolved into Digital Hangar: a comprehensive workspace that bridges the gap between multimodal data processing and traceable, structured outputs to speed evidence-based decision-making. 

Automated data triage

“Across government, teams are inundated with multimodal data, and it’s difficult to turn that volume into timely, reliable insight,” said Conor McSherry, associate director of AI engineering for Accenture Federal Services, while speaking at a recent Google Public Sector event at The Forge.

This challenge persists across many use cases, from law enforcement investigations and open-source intelligence (OSINT) collection to enterprise knowledge management. Digital Hangar addresses this bottleneck: 

• Rapid ingestion and processing: Users ingest case-related video, images, scanned documents, and reports. 
• Instant insights: The platform extracts structured signals, summarizes content, identifies key entities (people, places, organizations), and captures timestamps—all linked back to source evidence. 
• Automated briefing: Teams receive recurring summaries of what arrived in the last 24 hours, including volume, themes, and anomalies. 

Making connections between Insights

Digital Hangar transforms from a monitoring tool into an agent-enabled workspace where users can ask questions in natural language. AI agents built with Google’s Agent Development Kit (ADK) execute consistent analysis steps across the entire evidence set with full traceability.

Once data is ingested and key entities are identified and reviewed, users can also ask questions in a chat-based format, such as inquiring which entities appear most frequently. Through these queries, users can begin to make connections across the data and flag commonalities between incidents or reports.

While the user interacts, Digital Hangar works in the background to generate Discoveries—candidate correlations, recurring entities, and anomalies across Insights that investigators can then validate. 

“Sometimes something that is not important in a single piece of data is actually important when you look at all of your data combined,” McSherry said. “It’s not until you see it in 10 files that it’s now important and that’s where Discoveries come in.”  

AI-generated Insights power human decision-making

While automation plays a pivotal role, Digital Hangar is built upon a strict human-in-the-lead philosophy. The system generates timelines, summaries, and transcriptions, but final decision-making authority always remains with users. 

• Auditable: Authorized users can edit, annotate, or remove AI-generated findings. 
• Defensible: Strict traceability to underlying evidence ensures analysis holds up under scrutiny. 

As Digital Hangar continues to evolve and expand in capabilities, new iterations are aimed at leveraging additional data sources, moving beyond isolated deployments to connect directly to enterprise databases or various data management platforms—an effort already underway for the Digital Hangar team.

Innovation at mission speed

Digital Hangar’s AI capabilities leverage Gemini models for multimodal understanding of video, images and documents, and ADK to build and orchestrate agents that run repeatable analysis workflows. Within the Gemini Enterprise Agent Platform, ADK’s open-source agent development framework enables users to go from idea to easily configurable and reusable agents in minutes. With the support of Gemini and ADK, teams can enhance scalability, configure prompts, generate consistent structured outputs and iterate quickly as mission questions evolve.

Most critically, Digital Hangar eliminates the need to power through typical lags associated with government software innovation. Rather than building a bespoke, isolated government cloud, Google instead took the more complex route of accrediting its public cloud infrastructure at IL5 and FedRAMP High. This enables agencies to adopt modern cloud and AI capabilities within required security and compliance parameters, reducing the gap between commercial innovation and government deployment. 

Digital Hangar exemplifies the potential of the agentic era. Helping federal teams cut through evidence overload and move directly to actionable intelligence and by pairing multimodal AI with agent-driven workflows, it delivers structured, auditable insights while keeping human oversight at the center. 

As Stephen Hurley, managing director and platforms lead at Accenture Federal noted, “As the whole world is looking for a return on investment in AI, look no further than Accenture and Google Public Sector.” Digital Hangar demonstrates how the agentic era can modernize government operations and turn data chaos into mission value.

To learn more about how Accenture Federal Services and Google Public Sector are accelerating delivery of agentic AI solutions, request a visit to The Forge.