As Automated Driving (AD) and ADAS systems move closer to broader deployment, validation is realized as one of the industry’s hardest scaling challenges. Engineering teams no longer just need more miles or more simulations; they need the right mix of real-world data, synthetic scenarios, traceable requirements, and regulator-ready evidence to support safety cases and pre-deployment approvals.
This challenge is deepening further as regulatory expectations evolve across the EU, US, and global markets. In particular, the emerging 2026 UNECE ADS framework is expected to make auditable, scenario-based validation central to approval, increasing the need for engineering teams to produce traceable evidence across requirements, ODDs, safety assumptions, and simulation results. For OEMs, AV / ADAS, and Tier-1 suppliers, the change of rules, internal specifications drift and ODD expansion exerts more pressure on traversing the validation v-model from product requirements to completing the homologation. Manual validation workflows are increasingly difficult to maintain because every requirement, safety assumption, and ODD update can trigger the need for new scenarios, new evidence, and continuous re-validation.
That is where Deepen AI and Deontic’s partnership generates value for the ecosystem. Deepen AI brings the real-world data foundation: state-of-the-art multi-sensor calibration, data collection partners, high-quality pixel-level annotations, and encapsulating ML data operations that help teams build and validate perception and autonomy systems against safety standards. Deontic adds the synthetic and regulatory validation layer by using Generative AI to accelerate the creation of simulation assets from requirements, ODD assumptions, and compliance needs.
Deontic’s AI agent turns natural-language requirements into standards-based scenarios, including OpenSCENARIO and OpenDRIVE assets, while preserving traceability back to requirements and homologation evidence. In practical terms, Deontic helps teams move from “we need to validate this behavior” to “we have regulator-ready, simulation-ready scenarios” in a fraction of the time traditional V-model workflows require.
“Autonomous driving validation is becoming a data-to-evidence challenge. By combining Deepen’s real-world data intelligence with Deontic’s GenAI-driven scenario and homologation layer, we help engineering teams move faster from discovered coverage gaps to traceable, regulator-ready validation.“ – Stephen Lernout , Co-Founder and CEO of Deontic
Deepen AI starts from the real-world side of the validation lifecycle. OEMs, AV/ADAS companies, and Tier-1 suppliers generate massive volumes of driving data across sensors, geographies, road types, and operational domains. Deepen helps transform that raw data into usable ML validation inputs through annotation, calibration, data quality workflows, and safety-focused data operations.
The next challenge is knowing what the data actually covers and what it does not. As Deepen’s data curation capabilities identify rare events, edge cases, coverage gaps, and “needle in the haystack” driving situations, those insights can become direct inputs for Deontic. Missing coverage in a target ODD can be translated into new synthetic scenario requirements. A real-world event can be converted into structured simulation assets. A validation gap can become a parameterized scenario family.
This is Deontic’s role in the joint workflow: converting requirements, ODD constraints, regulatory expectations, and curated driving insights into ASAM-aligned simulation scenarios. Deontic’s GenAI scenario-engineering capabilities include natural-language-to-OpenSCENARIO/OpenDRIVE generation, parametric expansion across weather, actors, environment, and behavior variables, scenario libraries, and simulator integration. The result is a repeatable path from discovered data gaps to executable validation coverage.
Deontic is designed to sit where AD engineers already work: inside coding environments, CI/CD pipelines, requirements tools, simulation stacks, and homologation workflows. Through its agent, API, and MCP-based tooling, Deontic connects requirements, ODD constraints, scenario generation, simulation, and compliance monitoring into a single traceable loop.
This matters because end-to-end autonomy does not remove the need for scenario-based safety evaluation. In fact, when model internals become harder to interpret, the scenario layer becomes even more important. Deontic preserves auditability by converting requirements and ODD assumptions into testable, traceable scenario evidence. If VLMs and end-to-end models provide the semantic intelligence of autonomy, Deontic provides the validation intelligence layer that turns that intelligence into auditable proof.
Once Deontic-generated scenarios are simulated, Deepen’s triage and analysis capabilities can help close the loop. Failures, near misses, and unexpected behaviors can be analyzed, root-caused, and fed back into model development, dataset improvement, and scenario expansion. Together, Deepen and Deontic create a workflow where real-world data informs synthetic coverage, synthetic testing exposes system behavior, and analysis feeds the next cycle of autonomy improvement.
Deepen AI and Deontic are working together to reduce time to market for autonomous driving and ADAS programs by accelerating the full safety lifecycle: from data collection and ML data operations, to scenario generation, model validation, simulation, triage, and deployment readiness.
“The validation challenge demands a continuous, auditable workflow. With Deontic, we are closing the loop on AD/ADAS safety: transforming raw data into actionable insights, identifying critical gaps, and turning those gaps directly into standards-aligned validation evidence for homologation. This joint vision is the key to faster, safer deployment.” – Mohammad Musa , Co-Founder and CEO of Deepen AI
The joint vision-execution is a closed-loop validation workflow.
Deepen helps teams understand their real-world data, prepare it for training and validation, and identify the events and coverage gaps that matter most. Deontic helps turn those gaps into scenario-based validation assets that are traceable, standards-aligned, and ready for simulation and homologation workflows.
Looking ahead, this partnership becomes even more powerful as Deepen advances its data marketplace initiatives and Deontic expands ODD coverage and continuous homologation capabilities. Deontic’s roadmap points toward agents that monitor regulatory, legislative, and internal requirement changes; analyze their impact across features, ODDs, and scenario sets; and regenerate scenarios or prioritize regression testing to close the validation gap.
For AV/ADAS teams, the opportunity is clear: move faster without sacrificing safety, traceability, or regulatory readiness. By combining Deepen’s real-world data and validation expertise with Deontic’s GenAI-powered scenario generation and homologation intelligence, engineering teams can build a more scalable path from data to deployment.
To learn how Deepen AI and Deontic can support your validation, safety, and deployment workflows, schedule a quick technical discovery conversation with the teams.
Deepen AI can be reached at info@deepen.ai
Deontic can be reached at info@deontic.ai