Applied Intuition Copilot | AI chatbot

Challenges

Creating high-quality simulation scenarios manually is complex

Real-world maps can be difficult to procure, and building synthetic maps manually takes time

Writing SQL queries and building charts to understand test performance and drive logs can be time-consuming and technically challenging

Manually tagging drive logs is slow and expensive

Onboarding to complex new ADAS and AD simulation and data workflows can take non-technical users up to a week

Why Applied Intuition Copilot?

Applied Intuition Copilot features conversational interfaces and generative models that proactively assist teams and let them interact with simulation and data tools in the most intuitive and efficient way. It leverages large language models (LLMs), retrieval-augmented generation (RAG), and other specialized models like diffusion and perception foundation models to provide context-aware answers and accelerate critical tasks.

Generate simulation scenarios from simple text prompts to scale coverage across an operational design domain (ODD)

Create and edit maps using natural language prompts

Use natural language to generate SQL prompts to query and understand drive logs and test data

Tag and annotate drive log data automatically for search using perception foundation models

Rapidly onboard and learn workflows in Applied Intuition’s ADAS and AD development platform

Benefits

Maximize productivity

Interact with and learn to use Applied Intuition’s development platform intuitively using natural language. Search for information across documents and take action with simple prompts.

Accelerate development

Use simple text prompts and generative models to accelerate tasks like scenario creation, log tagging, and result analysis. Shorten development cycles and reduce operational costs.

Enhance system safety

Discover new edge cases and system performance insights hidden in log and test data by exploring and analyzing data with generative AI.

Key components

Simulation scenario generation

Create and edit simulation scenarios from natural language prompts, system requirements, or test cases. Build a scenario library up to 40x faster than with manual scenario creation tools, reducing multiple iterations and error-prone changes.

Generate scenarios in both OpenSCENARIO V2.0 and Applied Intuition’s scenario format. Further edit them using natural language, Applied Intuition’s scenario editor GUI, or their native scenario language.

Object Sim Sensor Sim Validation Toolset

Map generation and editing

Generate and edit maps using a simple prompt-based interface, and rapidly scale coverage across an ODD.

Utilize generated maps to run tests in Object Sim or feed them into Map Toolset’s procedural pipeline to generate 3D worlds for Sensor Sim.

Maps Toolset Object Sim Sensor Sim

Simulation test and performance analysis

Conversationally generate SQL prompts and build charts to understand a system’s performance. Reduce the time to create SQL queries from minutes to seconds.

Explore data from real-world and simulation tests, and review insights auto-generated by models to discover new edge cases.

Validation Toolset

Drive data auto-tagging and natural language search

Generate embeddings and tags for drive data using open-vocabulary perception foundation models.

Utilize generated metadata and natural language to find edge cases up to 20x faster. Link auto-tagged data to an ODD taxonomy for automated coverage analysis.

Data Explorer

Knowledge retrieval

Onboard rapidly and learn new workflows using conversational AI. Applied Intuition Copilot provides step‑by‑step tutorials, code snippets, summaries, and links to key documentation.

ADAS and AD development platform