Methodology
How we read the law for you.
Every line Obsidian delivers rests on an official text, at its own address, dated and checked before it reaches you. This page explains how we get there, and where our work ends and your experts' begins.
A companion for regulatory experts
You keep the judgment. We take the rest.
Regulatory and compliance experts are hired to interpret, to advise and to decide. In practice, much of their week goes elsewhere : checking regulators' websites, reading official journals, keeping registers up to date, sorting the changes that matter from the ones that do not.
Obsidian takes that work off their desk. We collect the sources, monitor them continuously, keep the regulatory database current and prepare a first reading of the impact for each entity. Your experts start from a structured, sourced answer, and spend their time on the part of the job only they can do.
What we handle
- Collection
- Every official source of your scope, its feeds and its journals : 10k+ sources in 150+ jurisdictions.
- Monitoring
- More than a million signals screened. Only what touches your entities reaches you.
- The regulatory database
- Each regulation kept as a record : its obligations, its deadlines, its status and its history.
- The first impact reading
- What changes, for which entity, by when, with the article it rests on.
What stays with your experts
- Interpretation
- They read our analysis with the official text one click away.
- Decision
- They confirm or dismiss each finding. Nothing is decided for them.
- Advice to the business
- They bring a sourced, defensible answer to management and to the board.
- Accountability
- Every decision is recorded : who, when, and on which basis.
Built with the profession
An engine designed with the people who do the job.
Obsidian was built on data and AI engineering principles, in collaboration with regulatory affairs and compliance experts. What a register must contain, what monitoring must leave out, what an auditor will ask to see : these rules were set by practitioners, and the engine applies them to every deliverable.
Artificial intelligence lets us read and cross-check at a scale no team could reach. It does not decide. Each output is held against its official text before delivery, and the final call always belongs to your experts.
The method
Five steps, for every line we deliver.
Source
We read the law where it is published.
Regulators, official journals and standards bodies, at their own address. Never a search result, a blog or a summary of a summary. If a text is not published by the authority that issues it, it is not a source.
Detect
Every signal screened, only yours kept.
Our watches read every new publication of 10k+ official sources and feeds. Each item is screened against your scope : what does not concern you is set aside, what does reaches you dated and sourced.
Structure
Each regulation becomes a record.
The text, its obligations and its deadlines, and its life cycle : proposed, adopted, in force, amended, repealed. Each record says whether a rule is law or only announced, so a draft is never mistaken for an obligation.
Assess
Read for each of your entities.
Each entity is assessed on its own activities and on your answers : applies, does not apply, or to confirm when information is missing. Every verdict states its reason and the article it rests on.
Verify
Nothing leaves before it is checked.
Every line is held against its official text before delivery. A statement that cannot be traced to its source is not delivered. Your experts then confirm or dismiss each finding, and each decision is recorded.
See the method applied to your own industry : a complete sample, on real regulations, by email within one business day.
Get your free sample Contact us