Introducing Dana: A New Way to Build Physical AI
Applied Intuition’s agentic platform dramatically accelerates how machines learn, improve, and get deployed in the real world.
Industries are entering a new era of intelligent machines.
Cars and trucks that drive themselves. Construction equipment that operates autonomously. Robots that continuously learn. Drones that swarm adaptively. And mines, ports, and factories that become smarter every day.
But how do you safely deploy intelligent machines at a scale never before imagined?
Today, the bottleneck is the development stack. Data lives across systems. Simulation, training, and validation often happen in different environments, and engineering teams are forced to move between legacy tools. Context is lost in handoffs. Lessons learned through years of development often become tribal knowledge that is difficult to scale.
We set out to change that.
Introducing Dana.
Dana is Applied Intuition’s agentic platform for developing physical AI applications, made for the complexities of building and deploying intelligent, safety-critical machines in the physical world.
For us, Dana isn’t just a new product. It’s a new way to build physical AI applications at scale, one that can accelerate the development of intelligent machines from sea to space and make our lives safer and more productive.
“Agents have transformed our digital world, and now we’re bringing that revolution to the physical one,” says Qasar Younis, co-founder and CEO of Applied Intuition. “Dana is a groundbreaking new platform that will help bring intelligence to a billion machines.”
Dana’s biggest advantage is the intelligence built into the platform.
Since 2017, Applied Intuition has worked on some of the hardest problems in physical AI alongside many of the world’s top OEMs and autonomy developers. Dana takes that deep domain experience and makes it available through intelligent agents and workflows in a single interface.
Developers can access Dana via natural language, APIs, SDKs, custom applications, and embedded interfaces, and it even integrates with enterprise systems and collaboration tools, like Slack and Jira. The result is a unified platform that helps companies build and deploy intelligent, safety-critical machines dramatically faster.
It gives developers an agentic platform that helps them move from a concept to a production deployment in rapid time.
“With Dana, we didn’t just build AI into existing products,” says Peter Ludwig, co-founder and CTO of Applied Intuition. “We took the infrastructure, tools, workflows, and expertise we’ve built over the last decade and made them agent-operable. That gives Dana the context and capabilities to help engineers solve real physical AI problems much more quickly and accurately, helping to expand the potential of intelligent machines across industries in new ways.”
Applied Intuition has been using Dana internally since last year, building and delivering solutions on the platform for long-standing customers across automotive, trucking, mining, and agriculture. Dana’s agent-driven workflows have reduced critical phases of vehicle development timelines from months to days in some cases. Applied Intuition has offered limited, early access to select customers, including Isuzu Motors, who is using the platform to accelerate L4 autonomy for its fleet of commercial trucks.
“We’ve been impressed by how Dana can streamline complex engineering workflows and accelerate development,” said Yasuhiro Yazawa, Director, Isuzu Motors Limited, Japan. “Dana gives our engineering teams greater confidence to develop, track and deploy safe autonomous-vehicle capabilities at a much faster pace.”
Another customer with early access is Komatsu, a leading manufacturer of heavy equipment for the construction, mining, and forestry industries.
“Applied Intuition has been a valuable technology partner as we continue advancing the digital capabilities that support the next generation of mining equipment and solutions,” says Peter Salditt, CEO, Komatsu Mining. “Dana represents another step forward, bringing intelligent, agentic capabilities into our engineering workflows to help our teams innovate faster, improve efficiency and ultimately create greater value for our customers' operations.”
Internally at Applied Intuition, we have replicated with far less code the functionality of complex products and applications that took years to build. Months of work was reduced to weeks or even days. Development cycles became 20x faster, with higher output quality. Engineers went from deployments once every few weeks to deploying five to ten times a day, on average.
The most ambitious example is our autonomy tools platform, which we have spent years building and refining. With Dana, we rebuilt a significant portion of its core functionality in six months in order to make it available with agentic capabilities.
A critical part of developing an end-to-end autonomy stack is building a production-grade flywheel: a continuously running pipeline that turns massive volumes of real-world and synthetic sensor data into better models and, ultimately, safer systems.
Building that flywheel is exceptionally difficult. It requires deep expertise across data, simulation, training, inference, evaluation, onboard software, cloud infrastructure, and deployment. Large teams can spend years building the underlying systems before they can begin improving them at scale.
Dana brings those capabilities together around three core stages: data, workflows, and insights.
Developers can use Dana for Physical AI to access and curate production-ready sensor data or ingest and validate their own. They can filter massive datasets for events of interest, map locations, or data-quality metrics, then prepare that data for training, inference, or simulation.
From there, Dana can orchestrate complex workflows across thousands or millions of jobs. Developers can run inference on autonomy models, evaluate regressions through open-loop replay, reconstruct real-world scenes for closed-loop neural simulation, generate adversarial behaviors with reinforcement learning agents, or use world models to create variations in weather, lighting, and environmental conditions.
Every result remains traceable, with full lineage across the workflow.
Dana then brings those results into a common metrics and analysis layer. Instead of manually inspecting thousands of runs, developers can compare autonomy stacks on safety-critical metrics, generate dashboards, query results, identify failures, and use specialized agents to determine what to do next.
In short, Dana keeps the flywheel moving—and the system learning—with every turn.
“The real breakthrough is the continuous development loop,” Ludwig says. “Dana can carry context from data to simulation to evaluation and back again. Every run produces information that can improve the next one, while maintaining the traceability and rigor required for safety-critical development.”
This is where Dana moves beyond a general-purpose AI assistant. It combines state-of-the-art models with the tools, infrastructure, data, and domain knowledge required to put those models to work on physical AI.
The same principle applies to other physical AI applications, like developing AI-defined vehicles.
Consider something as seemingly simple as a personalized welcome-lighting sequence for a passenger vehicle. Today, changing that feature can require work across requirements, architecture, code, multiple ECUs, simulation, hardware, and the vehicle itself.
Dana can connect that process into a single workflow.
It can read and evaluate requirements, reason across software architecture, work with code in reproducible cloud environments, validate changes in software-in-the-loop and hardware-in-the-loop environments, and ultimately test the same software on a real vehicle.
Just as important, Dana carries context across the process. Work that once moved across disconnected tools and specialized teams over months can, in some cases, be completed by a single developer, systems engineer, or product manager in an afternoon.
“We’ve spent the last decade building the infrastructure behind physical AI alongside some of the most advanced engineering organizations in the world,” Younis says. “Dana turns that experience into a platform that can scale.”
Dana is one of the most ambitious products we have ever built.
We didn’t design it for a single application or industry. We built it to work with all of them. Develop and validate end-to-end autonomy for cars and trucks. Build and operate autonomous fleets. Accelerate software-defined vehicle development. Orchestrate robots in industrial sites. Develop operating systems for mines, ports, and other complex environments.
In the coming weeks and months, we’ll be showing more of what Dana can do across autonomy, vehicle software development, fleet operations, and more.
Over the next decade, intelligence will be embedded in vehicles, robots, industrial equipment, factories, and machines operating throughout the physical world. And as those systems become more capable, the bottleneck will increasingly shift from creating intelligence to putting it to work—safely, reliably, and at scale.
That is the future Dana was built for.
Dana is a new way to build physical AI, one that moves at the speed of AI.
To see Dana in action and request more information, sign up here.
