HERE Technologies and AWS: Accelerating NDS.Live with agentic AI

The manual workflows that power map data development can become bottlenecks quickly. Teams spend countless hours on repetitive tasks—validating data quality, documenting compliance requirements, coordinating across departments—when they could focus on innovation. HERE Technologies faced this challenge with NDS.Live, their next-generation map data platform. To address it, they partnered with AWS to implement an orchestrated multi-agent AI system that automates these workflows while maintaining the strict traceability automotive standards demand. The result: a largely automated process that replaced fragile manual work with consistent, scalable operations.

At its core, the solution uses Kiro, an agentic orchestration framework that coordinates multiple specialized AI agents to handle different aspects of map data development. Think of it like an automated team: one agent handles data validation, another manages compliance documentation, and another coordinates handoffs between teams. These agents don’t work in isolation—Kiro orchestrates them through a CLI tool and a three-layer memory system. The first layer holds immediate task context, the second maintains longer-term project state, and the third stores institutional knowledge about processes and standards. This architecture means agents can reason across their work, learn from previous tasks, and maintain consistency across large, complex projects. Importantly, the system captures ASPICE traceability automatically—a critical requirement in automotive development where regulatory audits demand proof that every decision was documented and justified.

The practical impact is substantial. Before implementing agentic AI, HERE’s teams struggled with manual coordination: sending emails to confirm requirements, manually filling out compliance matrices, and hand-checking data against standards. These processes were error-prone and consumed engineering time that could have gone toward improving map accuracy. Now, agents handle these tasks automatically, with human teams reviewing and approving the results. This shift lets engineers focus on higher-value work—analyzing complex edge cases, optimizing data quality, and designing new features—while maintaining the documentation rigor that automotive products require.

For teams building similar systems, this approach offers a practical blueprint. Start by identifying workflows that are repetitive but require consistent, auditable execution—map data validation, compliance documentation, and cross-functional coordination are classic examples. Then consider whether an orchestrated agent system makes sense for your use case. You’ll need to handle memory carefully so agents can access the context they need, define clear handoff points between agents, and ensure traceability is captured automatically rather than bolted on afterward. HERE’s experience shows that when these elements align, agentic AI can transform labor-intensive processes into scalable, maintainable workflows that grow with your business.