HERE Technologies and Wherobots partner to give AI agents geospatial intelligence

October 14, 2026 – Munich, Germany – Wherobots and HERE Technologies, the world’s leading mapping and location data company, have formed a strategic partnership to enable AI agents to understand and make precise, deterministic decisions about the physical world.

Initially, Wherobots and HERE are integrating HERE Location Reasoning (HLR) with the Wherobots MCP (Model Context Protocol) server. Customers of this integrated solution can now build AI agents that combine complex and large-scale physical world analysis with precise geospatial intelligence in a single workflow. These include use cases such as last-mile delivery and logistics planning, infrastructure risk computation, dispatch design and resource allocation, site selection, property and casualty underwriting, and more.

AI has the potential to positively transform how and where assets are allocated and move. But to achieve this, AI must be capable of effectively utilizing the data about the physical world. Currently, AI agents heavily rely on external tools to generate the context required to make decisions, but, because of the preference of AI models to select free and open tools that they can easily harness, the teams that utilize them frequently start with solutions that seem initially robust, but are often flawed. They can’t analyze the data across spatial modes at scale, and often require bespoke and complex builds to deliver answers that are only partially reliable.

Together, Wherobots and HERE bring the tools designed to enable AI agents to understand the physical world through large, diverse datasets and deterministic guidance, enabling a new tier of optimization and innovation. Through an MCP integration, AI agents turn on high-performance inference and analytics on multiple types of spatial and non-spatial data at various scales within Wherobots, and join this with precise geospatial intelligence offered by HLR. The result is a high-value agentic workflow, with the capability and precision that can unblock innovation and resource optimization across industries and use cases.

Quotes

"AI is moving from generating content to making decisions and taking action in the physical world. That requires more than a language model. It requires a trusted understanding of where things are, what rules and conditions apply, and what actions are possible in the physical world. The partnership between HERE and Wherobots gives developers a simple way to bring that geospatial intelligence into the systems they are already building."

– Mahesh Shah, Senior Vice President of Enterprise & AI Solutions at HERE Technologies

"Agents increasingly answer questions, complete tasks, and run workflows on our behalf, and many of those tasks require deterministic answers about the physical world. Adding HERE Location Reasoning to the Wherobots context engine for physical world data gives people and agents that precision in one workflow. Leading technology and data providers have a responsibility to make that possible."

– Mo Sarwat, CEO, Wherobots

Why we're building this

Wherobots and HERE have been working on this problem from complementary directions for some time. Wherobots processes satellite imagery, mobility, weather, risk and operational data at regional to planetary scale to produce forecasts, scores and enriched datasets. HERE Location Reasoning provides AI-native access to trusted HERE geospatial intelligence, including live traffic, geocoding, place discovery, passenger and commercial vehicle route optimization and EV charging intelligence.

The future of agentic workflows need both. The partnership between the two companies makes it easy for agents to process large scale and historical physical world datasets, as well as inform decisions with live, on the ground intelligence, so that AI can accelerate and increase the accuracy of their location-aware development cycles.

How it works

An AI agent connects to one endpoint, the Wherobots MCP server. Wherobots sends each sub-task to the engine built for it:

• Wherobots: large-scale spatial joins, imagery and mobility processing, forecasts, risk surfaces, and answer refinement.
• HLR: geocoding, search, routing, isochrones, place semantics, and constraint-aware decisions.

What agents can do

Each workflow below runs as a live demo app in the Havasu Knowledge Lake.

• Wildfire evacuation routes. Wherobots scores all 357,554 transmission spans in California for wildfire exposure. A span is the stretch of line between two towers. Wherobots then measures how much of each evacuation route runs beside high-risk spans. HLR computes the routes out of each community, times them at rush hour, and maps how far a car gets in 15, 30 and 60 minutes. Demo: Mount Hermon is a forest town in the Santa Cruz Mountains. All 21 HLR routes from the town to four destinations run beside high-risk spans for 0.79 to 2.04 miles. The least exposed route to the coast takes more than twice as long as the fastest. Read the analysis · Open in Havasu
• Property and casualty underwriting. HLR geocodes and validates each address at rooftop and place-level precision. Wherobots scores the property against flood, wildfire, hail, wind, and earth-movement risk surfaces, plus recent imagery and ML damage predictions, then adds HLR's emergency-response accessibility. The agent returns one underwriting decision. Users can also ask in plain language, such as "the top 10 properties at risk within 10 miles of downtown Denver," and get ranked results on a map. Open in Havasu
• Last-mile network design. Wherobots screens a metro for drone-nest and depot sites against demand, airspace, terrain, and buildable land. HLR travel-time reach then ranks each candidate by service-area coverage under real road and airspace constraints. Demo: Texas drone nest siting. Open in Havasu
• Telecom capacity and coverage. Wherobots combines speed tests, tower and infrastructure inventory, terrain, and tree canopy to separate capacity-limited from coverage-limited service and find underserved areas building by building. HLR routes field crews to the sites with the largest reliability impact. Demo: California signal quality. Open in Havasu
• Predictive logistics routing. Wherobots combines global weather forecasts with historical traffic and road context to predict tomorrow's conditions across every metro a carrier serves. HLR plans truck-legal, constraint-aware routes against that forecast to minimize cost. Demo: Route Sentinel scores freight corridors against near-real-time wildfire, smoke, and red-flag hazards and proposes cross-modal reroutes. Open in Havasu
• Data center site selection, continent to address. Wherobots screens every viable parcel across a continent against demand, demographics, mobility, and competition. HLR evaluates each finalist with drive-time isochrones, place semantics, and junction-level access quality. Demo: the National Suitability explorer screens parcels across Australia down to address-level fit. Open in Havasu
• Utility wildfire risk and crew dispatch. Wherobots scores every transmission span for ignition risk from imagery, vegetation, and weather. HLR routes inspection crews to the right access points under truck access rules and seasonal closures. Demo: the California grid wildfire app scores every span statewide across fuels, NDVI vegetation, terrain, and fire weather, with current conditions from VIIRS and GOES thermal detections. Open in Havasu

Outputs can be delivered to any cloud storage or lakehouse.

These examples show how the Wherobots integration with HLR can support broader real-world AI use cases across logistics, infrastructure, insurance, site selection, mobility, field operations asset deployment and customer support, where AI systems need reliable spatial context at scale, with precise intelligence to support operational decisions.

Availability

Within the Wherobots integration, HLR is available through Wherobots Spatial AI Tools and MCP Server, with support for the Wherobots first-party agent coming soon. To explore the integration, documentation, examples and reference workflows, visit wherobots.com.

About Wherobots

Wherobots is the context engine for physical world data. It is the data and infrastructure layer that enables processing of small to large scale physical world data, geometries such as buildings, waterways, road networks, as well as imagery from satellites, drones or airplanes. Existing AI infrastructure was trained on text, code, and traditional database tooling, but the most consequential enterprise decisions often deal with assets and infrastructure in the physical world: supply chains, properties, geopolitical risk, and infrastructure.

Built by the original creators of Apache Sedona, Wherobots lets teams process imagery at planetary scale, run precise spatial joins across billions of geometries, deploy computer vision models on earth data, and give AI agents persistent memory of everything their organization has built within the physical world.

Apache Sedona and Wherobots are used by some of the most forward thinking companies to optimize their products in the realities of the physical world and stay ahead of an evolving planet including: Amazon, Nearmap, Ordnance Survey, Allstate, Uber, T-Mobile, Zillow, Oak Ridge National Laboratory, J.B. Hunt, Land O’ Lakes, and more.

Contact:

Lindy PR
michael@lindypr.net