Compliance teams frequently face overwhelming regulatory shifts like those imposed by MiFID II or EMIR Refit. For many firms, understanding exactly how a new mandate impacts day-to-day operations can feel overwhelming – unless, of course, you can assess the operational impacts and immediately trace those mandates back to the source text itself.
Since its founding in 2012, Droit has methodically addressed this challenge head-on, operationalizing hundreds of global regulatory mandates – Dodd-Frank, MiFID II and EMIR Refit among them – into structured, actionable decision logic.Instead of treating regulations as abstract legal texts, Droit transforms them into clear, structured data sets, a method central to its Adept platform. Within Adept, compliance professionals don’t wrestle with interpretive ambiguity; they navigate intuitive decision trees linked directly to annotated source regulations.
RegTech Insight recently caught up with Brock Arnason, CEO and Founder at Droit to dig deeper into the foundational principles that power the company’s Adept platform
Turning Law into Actionable Data
At the core of Adept’s innovative approach is a philosophy clearly articulated by Arnason, who emphasizes that the transformation of legal and regulatory texts into structured, actionable information is fundamental to Droit’s mission. In Arnason’s words:
“We’re trying to get people to buy into a vision of treating the law as data and to subscribe to a data service that defines the best-practice, consensus interpretation of the law. Our infrastructure draws on that data service to make decisions and to help explain those decisions with complete transparency and traceability to the underlying text,” he says.
Central to this vision is the systematic normalization of legal texts. Droit meticulously monitors regulatory documents and converts them into a standardized, structured representation that remains clearly versioned over time. Arnason specifically highlights the importance of maintaining this structured and standardized regulatory content, referring to it explicitly as “data artifact number one.” This artifact is foundational, providing a consistent and unambiguous reference point for all downstream regulatory decisions and compliance activities.
The resulting structured data forms the backbone of Adept’s platform. Users engage with this content through a specialized web application – the Digital Library – that offers intuitive, user-friendly features. Within this library, compliance professionals can easily view differences between multiple versions of a regulation, conduct targeted searches, and access comprehensive annotations layered onto the regulatory text. These annotations, Arnason explains, act as critical “connective tissue,” linking precise paragraphs of source texts directly to the platform’s underlying decision models.
By converting ambiguous, text-based regulations into standardized data, Adept enables firms to achieve enhanced transparency and precision in their compliance processes. Regulatory mandates are transformed from rules to interpret into structured logic that can directly drive decisions. This method ensures clarity in both interpretation and application, enabling institutions to rapidly and confidently respond to regulatory requirements.
Visualizing Compliance: The Power of Structured Logic
The Adept platform is based on symbolic logic AI, a structured approach with roots in classical expert systems from earlier generations of artificial intelligence. This methodology has enabled Droit to build compliance models that are fully transparent, explainable, and robust enough to withstand rigorous regulatory scrutiny. By translating complex legal text into clearly structured logic, Adept ensures that every compliance decision made within the platform can be traced, justified, and consistently replicated.Arnason described this as “a first order logic representation – something that you can express or render as a human readable decision tree that operates on input data elements.” This structured logic representation means that compliance officers and operational teams no longer need to wade through dense regulatory documents manually. Instead, they are guided through clearly delineated decision trees, which explicitly detail the conditions and exceptions embedded within regulatory texts.
The user-friendly nature of Adept’s decision logic is further enhanced through two dedicated web applications: the Digital Library and the Logic Viewer. The Digital Library allows users to review, compare, and annotate different regulatory texts, capturing versioned changes clearly. These annotations serve as links from the original regulatory texts to Adept’s structured decision models.
The Logic Viewer, described by Arnason as enabling users to visually “trace regulatory decisions back to their original source texts,” provides compliance teams with a clear, actionable interface. Users can quickly pinpoint exactly why a certain regulatory obligation applies or does not apply to a specific scenario. For example, querying the source of a particular rule to check your exemption status and being directed to the pin site of the source text that has the paragraph that actually carves out that exemption.
By structuring regulatory compliance into explicit logic models linked directly back to source documents, Adept not only reduces ambiguity but also dramatically improves the speed and accuracy of compliance decision-making. This approach inherently addresses critical regulatory demands for transparency, auditability, and version control – attributes that regulators expect financial firms to demonstrate clearly and explicitly.
Limitations of Traditional Symbolic Logic Approaches
While symbolic logic significantly enhances regulatory precision and transparency, its deterministic nature comes with inherent limitations, particularly when dealing with uncertainty or predictive analyses.The structure of symbolic logic – where decisions are represented in explicit, deterministic, mathematically rigorous terms – means it is less suited to handling queries that require probabilistic inference or generalized prediction.
In describing these constraints, Arnason emphasizes a critical point of distinction: “You can’t ask a generalized question of our system and then have it give you a generalized answer. You have to ask it a question in a formalized way with inputs, and it will give you a very precise answer.” Symbolic logic does not inherently learn from vast data sets or generalize beyond explicitly encoded logic, making it unsuitable for broader predictive analysis tasks or open-ended questions.
Given these constraints, Arnason outlines how Droit is exploring the integration of modern AI methodologies – particularly Large Language Models (LLMs) – as complementary technologies to enhance Adept’s capabilities.
Droit’s experimentation with LLMs includes improving operational efficiency in generating explanatory text to support compliance decisions. Initial testing revealed that, GenAI alone struggled with both accuracy at 40-60% and performed even worse with consistent explanations. However, when the LLMs were given structured guidance and context from Adept’s decision logic, the results became “shockingly good.”
By pairing symbolic logic with the capabilities of LLMs, Droit looks to close the gap between precise, deterministic compliance decision-making and flexible, context-aware explanatory capabilities.
In recognizing the strengths and limitations of each technology, Droit is actively exploring how generative AI can augment and enrich its existing structured compliance framework.
Managing Regulatory Change
Adept excels at managing regulatory changes precisely because of its structured data and logic-driven model. Droit currently supports hundreds of regulatory mandates, enabling financial firms to seamlessly track, interpret, and implement regulatory updates. As Arnason notes, the platform significantly reduces the complexity associated with understanding and integrating these changes. For example, when regulations evolve, Adept provides detailed comparative analyses, clearly highlighting differences between regulatory versions. By generating detailed before-and-after scenarios, Adept simplifies the complexities inherent in regulatory adjustments, allowing firms to proactively address the specific changes required in their data collection and compliance processes.
Moreover, Adept’s structured approach empowers institutions to accurately assess how regulatory amendments will directly impact their business portfolios. According to Arnason, this portfolio-level insight is highly valuable, as it enables a precise understanding of regulatory implications across complex operational environments. He elaborates: “You can run this through our decision engine on a before and after basis, identify… this fraction of your decisions have changed… We can even trace which parts of the legislative change are identified with those particular change decisions.”
By clearly identifying and mapping the exact portions of regulatory text driving changes in compliance outcomes, Adept offers a powerful tool for managing operational risks associated with regulatory shifts.
Additionally, the frequency and method of Adept’s regulatory updates underscore the platform’s agility and responsiveness. Adept delivers continuous incremental updates directly through data releases, bypassing the often-cumbersome software deployment cycles that typically accompany regulatory technology updates. Arnason emphasizes this efficiency, noting: “We’re doing more than one release per day on average. And those are data releases.” This continuous data release cycle ensures that firms can rapidly adapt to regulatory changes as they occur, keeping compliance frameworks consistently up to date without disruption or delays.
Droit’s vision for further leveraging AI within compliance processes remains thoughtfully ambitious, targeting specific, clearly defined enhancements. One area under active exploration involves leveraging AI to automate test-case generation – an often tedious but critical aspect of compliance engineering. Automating this process through advanced AI promises notable efficiency gains internally, transforming a traditionally manual and repetitive task into one supported by intelligent automation.
Another intriguing opportunity lies in semantic interoperability between distinct data models and the potential of newer and more advanced large language models (LLMs) to assist in creating detailed analysis and mappings between different client and Droit-specific data schemas. The advantage here is significant, as it would facilitate smoother integration, reduce onboarding complexities, and enhance overall efficiency when aligning compliance data across systems.
Nevertheless, Droit is distinctly measured in its approach to these innovations. Arnason underscores this careful stance, emphasizing a structured, incremental approach rather than rapid adoption. He explains, “We are very cautious in how we roll these things out. We have to test extensively, be very confident in them. But this is all in our AI roadmap.”
Subscribe to our newsletter
