Answers to the questions practitioners, technologists and legal operations teams most commonly ask: about what the taxonomy is, how to use it and how to contribute.
noslegal is a community of people working in law with an ethos of "open source legal for all of us."
Our primary product so far has been a shared practical taxonomy for legal work: a standardised vocabulary and classification system, with concepts subdivided in ways which reflect practical needs.
The goal is practical interoperability: when a law firm, in-house team, tech vendor and analytics platform all use the same terms with the same meanings, data flows cleanly between them and comparisons become meaningful.
The project is community-led but with careful governance. Anyone can use it, free of charge. The licence conditions are minimal: attribution, and no use of the noslegal name to imply endorsement.
The taxonomy is a structured set of eight classification frameworks (called facets) that together describe the full landscape of legal work. Each facet covers a different dimension: what kind of legal work is being done, what area of law is involved and so on. They are designed to be used together but each can also stand alone.
First, it is genuinely international. It was not designed for a single jurisdiction and then adapted for others. Its concepts are rigorously jurisdiction-neutral, with jurisdiction-specific content to be handled through separate modules. We've provided a few of these already (called extension packs) and welcome suggestions for more. You're also free to extend it with other specialised taxonomies (e.g. for particular countries or business contexts) or to add your own concepts.
Second, it is simple by design. The structure is deliberately straightforward and the content is lean. Both are easy to grasp, and in v4 we've also put great effort into web design and more detailed guidance on how to use it.
Third, it is carefully faceted to allow concepts to be combined in ways that create powerful descriptiveness without concept-proliferation.
Fourth, it is designed around legal work as it is actually performed. For example, our Work types facet classifies work by its purpose, not just its technical legal subject matter and not just from a law firm perspective (e.g. specific legal services). Work types are structured to support planning and delivery, with detailed process elements captured separately. This allows organisations to understand and capture data relevant to getting things done.
Fifth, it is focused only on four practical areas of need (delivery, pipeline, knowledge and people) across three types of organisation (legal service providers, their clients and software providers). It does not skew to just one or two organisation types, nor does it extend to academic or doctrinal use cases (which tend in practice to lead down local and technical doctrinal paths).
Sixth, it is licensed on permissive open source terms, so you can freely study, use and modify it. You can even redistribute your modified versions if you want.
Seventh, recognising that adoption of standards can be challenging for organisations with embedded legacy taxonomies, we have emphasised in the taxonomy design and guidance its potential for use as a mapping or translation layer between existing classification structures, so as to increase interoperability across systems, datasets and knowledge sources.
Start with the taxonomy landing page. Browse the eight facets to get an idea of what they cover. Then read the guide: its six chapters can be read from the taxonomy landing page as well, or you can download it as a PDF. Refer to these FAQ or get in touch if you have any other questions or feedback!
It depends on how you see the value and risk of your legal data now and in future.
Even purely internally-facing data adequacy is hard to achieve in an organisation of any complexity. If you have more than one system, more than one team, and more than one taxonomy that has evolved over time without coordination, that's already a set of challenges. The question is not just whether your classification works in isolation but whether it works consistently across your own organisation: between groups, between offices, between the system where matters are opened and the system where costs are recorded. Inconsistency at those internal joins is the most common source of data that cannot be used for the purposes it was collected for.
The costs of inadequate classification also show up when sharing data with other organisations. For example, as a legal department when you instruct external law firms and want to combine their data with yours to improve how you handle matters of particular types. When you benchmark spend and outcomes.
These problems are becoming more consequential with the opportunities and risks which AI now offers. Legal data that sits in siloed, ineffectively classified systems was always suboptimal. It is increasingly a strategic liability. AI tools that assist with matter planning, spend analysis, resourcing and outcome prediction depend on consistent, well-structured input. "Garbage in, garbage out" applies with particular force when the model has no way to know that two things with different names mean the same thing, or that two things called the same mean different things. Organisations that invest in classification discipline now will be better placed to realise the value of those tools as they mature. Those that do not will find that inconsistent historical data limits what they can do, and that cleaning it up retrospectively is expensive.
noslegal does not solve a problem you do not have. It solves problems you probably already have, and that will matter more over time.
Ask knowledgeable lawyers what's actually blocking meaningful AI adoption in their legal department or law firm, and lack of structured data comes up near the top of the list. AI doesn't reduce the need for a taxonomy. Quite the opposite: it increases the value of having a good, well implemented taxonomy and magnifies the problems of having a problematic or poorly-implemented one.
Modern AI tools are good at retrieval, summarisation, drafting and extraction from unstructured text. They are much less reliable at consistent classification at scale. Ask an LLM to classify the same matter twice with slightly different prompts and you may well get different answers; ask it to count "our M&A matters" and the result depends on the prompt, the model and what else is in context. For anything requiring aggregation across a corpus - analytics, benchmarking, pricing, resourcing - that variability is highly problematic. A taxonomy gives you stable, deterministic labels that persist once applied. AI can help apply them, but you should be applying labels from a defined scheme rather than inventing them on the fly.
AI does not remove the need for shared standards: it raises their value, because translation layers between organisations work better when both ends share stable referents.
AI and taxonomy are complements, not substitutes. Structured inputs make AI outputs more reliable; AI tools make classification, mapping and extension cheaper than they used to be. The interesting question is not whether to keep using a taxonomy now that AI exists, but how to design the two together.
The core taxonomy is developed and maintained by a working group of legal professionals, pricing specialists, legal operations practitioners and technologists. Decisions on what goes into each release are made through a working group and consultation process.
The project is not owned by or affiliated with any individual firm, vendor or professional body. That independence is deliberate: the taxonomy is only useful as a shared standard if it is genuinely neutral.
Yes, it's free in all senses: you can do whatever you like with it, and there's no payment required. It is released under the Apache 2.0 licence, which permits free use, modification and distribution (including in commercial products) so long as you include the required attribution notice and do not use the noslegal name or logo to imply endorsement without permission.
The project is sustained by the in-kind contributions of its working group members and the organisations that support them. There is no subscription, licence fee or commercial tier.
You can absolutely pick and choose. But they share a common structure and are designed to work together, supplementing each other and not duplicating. For example, the Roles facet indicates which roles are relevant to particular Work types and Areas of law. So if you do pick and choose, take into account the extra design work you'll need to do in that regard. Also, updating later on to use the latest additions to noslegal is easier the less you depart from the standard.
The full taxonomy is available at noslegal.org. As at early May 2026, it is published in structured Excel format. The interactive landing pages here provide a navigable, human-readable overview of each facet's content and structure.
Yes. Please do! That is one of the mechanisms by which the standard achieves reach. The Apache 2.0 licence permits commercial use. If you integrate noslegal into a product, please include the attribution notice in your documentation and, where practical, link users back to noslegal.org so they can access the authoritative source.
A mapping exercise typically has three steps: (1) collate your current codes and the concepts they represent; (2) match each to the nearest noslegal equivalent at the appropriate level of granularity; (3) decide how overlaps and gaps are to be addressed, bearing in mind the benefits of using noslegal concepts for interoperability.
The noslegal spreadsheets include definitions and alternative names which are useful for this purpose. Using an AI tool can greatly speed up the mapping exercise nowadays but make sure to have appropriately knowledgeable humans check the mappings with your organisation's realities in mind.
The taxonomy is published in English. For multilingual deployments, the recommended approach is to use the noslegal code identifiers (which are language-neutral) as the stable internal reference layer, and maintain a local-language label set on top of them. The Places facet includes a dedicated extension pack for official UN non-English place names.
If you'd like to help us develop an authoritative version of noslegal in a particular language, please get in touch!
Version 4 of the taxonomy has initially been published (May 2026) in Excel format, with consistent column headers and code schemes designed to support automated ingestion. JSON and CSV versions are coming shortly on GitHub.
Major versions follow a formal consultation process and have been released on a roughly annual cadence to date. Minor updates (bug fixes, improvements and minor extensions) are released occasionally. New releases are announced at noslegal.org and via the GitHub repository.
Feedback, questions and suggestions can all be submitted via the 'Provide feedback' link in the navigation. If you want to cite noslegal in published work, the recommended format is: noslegal taxonomy v[version number], released under the Apache 2.0 licence. Available at noslegal.org.
There are two basic ways to use noslegal. One is to replace all or some of your existing taxonomy when you assess that noslegal is better. In doing so, you can extend noslegal to address further details relevant to your organisation. The second way is to use it as a mapping layer to join up different taxonomies within your organisation or between your organisation and others (e.g. your legal services providers or clients). Mapping documents the relationship between your existing concepts and the corresponding noslegal concepts. It is often an approximation rather than a perfect translation, but even an imperfect mapping consistently applied allows data classified under different taxonomies to be understood and combined.
To make rapid progress, the most important level to map to is often the second level: concepts at this level represent the core interoperable layer. Every departure from second-tier alignment generates a downstream cost to manage. The layer above it gathers these concepts together, and the layer(s) below it break down further details.
Where clients, law firms, and technology providers describe legal work using a mapped set of concepts in this way, it becomes easier to exchange information, benchmark performance, and integrate workflows across organisational boundaries. These benefits compound over time: an organisation that invests in alignment now, even partially, will be better placed to realise them as adoption widens. If your existing concepts lack explicit definitions, this is a practical opportunity to adopt noslegal definitions where possible. If you have your own definitions, consider how well they are applied in practice and whether they offer any material advantage over the noslegal equivalents. If not, alignment with noslegal is likely to be a good choice.
Enablement is a residual category. It applies to work that does not fall within a more specifically defined work type category.
The practical test is therefore to work through the other categories first. Is the matter concerned with resolving a dispute over legal rights or remedies? Dispute resolution. Establishing or amending a legal relationship or status? Transaction. Formal insolvency, wind-up of an estate, or separation of a relationship? Dissolution and restructuring. Constructing or materially altering a physical or intangible asset? Development. Structured fact-finding or official proceedings? Investigative process. And so on. Enablement comes into play when none of these fits: ongoing advisory support, day-to-day legal backing for a business function, internal initiatives, authority responses, and residual representation and advice all live here. Its ten types cover a wide range of situations. The unifying feature is that they fall outside any more specifically characterised work type but nevertheless enable an organisation or person to fulfil their objectives.
Work that is primarily advisory but falls within a more specific category (for example, advising a client through an official investigation) should be classified under that specific category.
We've concluded that legal operations is not a work type in the same sense as the others. It is better modelled as a cross-cutting approach that can be brought to bear across different work types. Treating it as a discrete work type alongside those categories was not useful. This means that in practice legal operations activity will typically be classified under whichever work type best describes what is actually being done: a compliance framework design project falls under GRC and policy; a post-merger integration programme falls under Enablement; a litigation spend benchmarking exercise would sit in Dispute Resolution. The legal operations dimension can be captured through topic tags or organisation-specific fields rather than at work type level.
We recognise that this may leave some people wanting more explicit guidance on how to represent legal operations consistently across classifications. We are planning a short, focused consultation exercise later in 2026 to consider how best to address this: whether through extended guidance, a dedicated extension pack, a definition in the process elements glossary or in other ways.
The most likely classification is within Enablement, specifically as Organisational initiatives (wkt-enb-7). The definition of that type covers legal support for the implementation of major internal programmes, including organisational change and business transformation initiatives. A legal operations capacity-building project (building legal operations capability, improving processes, implementing new tooling) would typically fall within that scope.
The caveat in the definition is worth noting: Organisational initiatives explicitly excludes projects that can be categorised in other work types. So if the capacity-building work is primarily a compliance programme, it would belong in GRC and policy (wkt-grc-3, Risk and compliance). If it is primarily a training and guidance exercise, Advice (residual) (wkt-enb-10) may be more appropriate. If it involves procuring and implementing technology, Organisational initiatives remains the better fit.
As with any classification decision, the question to ask is: what is the primary purpose of this project as a whole? Classification should follow from that purpose rather than from incidental features of how the work is being done.
Governance, risk and compliance. While we have generally avoided initialisms in category names, we concluded after consultation that this one is widely understood and used by people working in the field and can easily be found by anyone unfamiliar with it. Brevity in this case was more important. In implementation, you are free to use a fuller phrase as a synonym if you wish, as the unique noslegal code will identify it as the same concept.
noslegal does not impose a single classification scheme on everything about a matter. Data structures in practice typically include two types of field: fields that draw on the noslegal taxonomy (work type, area of law, sector, places, roles), and other fields that capture organisation-specific context using your own terms: practice group, business unit, internal client, cost centre, strategic initiative, product line, or whatever dimensions matter to your organisation. These are not alternatives. They can sit alongside each other in the same record.
The Sectors facet adds a further dimension worth noting. The same matter may legitimately be classified by matter sector (the sector most relevant to the legal work itself) and client sector (the industry or industries in which the client operates) as separate data points. These will sometimes differ: if an energy company instructs you on a real estate acquisition, classifying the matter in the real estate sector and the client in the energy sector is both accurate and useful for different purposes. Field design should make this distinction explicit.
The practical principle is that fields should be designed around the questions the data needs to answer. noslegal provides the shared conceptual foundation; your organisation's own business taxonomy and your field design can provide the additional business context.
Within Areas of law, the primary classification is by issues such as public law, land law and family law with indigenous law aspects being addressed as topics within these. If you are interested in developing an extension pack in this area either with international application or for a particular country, please get in touch.
In areas of law terms, the most relevant classification will depend on what the matter concerns. General government law within Sectoral laws > Public sector and NGO (law-sec-pub-6) explicitly includes municipal law as a topic and covers local government organisation, powers and decision-making. Planning law (law-lnd-2) covers land use decisions including municipal planning, zoning and development consent. Public law > Constitutional and administrative law (law-pub-1) or Public law > Rights law (law-pub-2) may be relevant if the matter concerns the legality of a municipal body's decisions or their impact on individual rights.
Work type will follow from the nature of the matter. A challenge to a municipal decision is most likely Litigation (wkt-dsp-2) or, if there is an administrative appeals route, Enablement > Representation (residual) (wkt-enb-9). Advising on compliance with a municipal regulatory requirement may be GRC and policy (wkt-grc-2 or wkt-grc-3). A contract with a municipality is a Transaction. Classification in the Public sector and NGO sector (sec-pub) is appropriate where the municipal body is itself the client.
The noslegal taxonomy classifies by the process that is actually underway or clearly expected, not by the process that might eventually happen. Classification should therefore follow the current state of the matter and be updated as the position becomes clearer. If formal proceedings have been commenced (in a court, tribunal, or under arbitral rules), the classification is clear: Litigation (wkt-dsp-2) or Arbitration (wkt-dsp-3) as appropriate. If the matter is still in negotiation without any formal process underway, Consensual resolution (wkt-dsp-1) may be used, but its definition should be noted: it applies to a free-standing process aimed at agreed resolution without any process already underway to impose a binding decision. It is not the right classification simply because parties have not yet gone to court.
Many matters begin with informal negotiation before any formal process and it is not always clear at the outset which route will be taken. In those circumstances, the best approach is either to make an initial classification and revise it as the position develops, or to subdivide the matter as the nature of the process changes, depending on the nature of your organisation's workload and data needs, policy on this is something to define and enforce. Materiality limits should be considered for defining when to reclassify and when to subdivide.
The question captures a genuine initial surprise, but the underlying logic is coherent.
The Dissolution and restructuring category (wkt-rsl) is defined by a shared purpose: winding up or restructuring the affairs of an organisation or individual, solvent or insolvent, alive or dead. What the types in this category have in common is not their legal subject matter but their process character. All of them involve bringing something to a close or fundamentally reorganising it, gathering and valuing what exists, resolving claims, and distributing assets or rights to those entitled to them.
Administering a deceased estate involves exactly that process: obtaining authority to act, identifying and valuing estate assets, settling debts and liabilities, and distributing the remainder to beneficiaries. Handling a divorce or separation involves dividing shared assets and resolving financial claims between parties. Formal insolvency and solvent reorganisation involve much the same thing at an organisational level. These are not superficially similar: they draw on the same process vocabulary and generate similar data needs for planning, pricing and experience management.
The alternative (placing deceased estates in Personal affairs and divorces in Family law) would make sense if the area of law were the organising principle. But noslegal's Work types facet organises by purpose, not by area of law, which is captured separately in its own facet. When a lawyer plans, prices and manages a deceased estate or a divorce, the relevant experience for comparison and knowledge purposes has as much in common with solvent restructuring work as with other family law or personal affairs matters. Keeping them together reflects that reality.
The areas of law facet handles the legal subject matter: Succession and estate administration (law-paf-2) for estates, Divorce and separation (law-fam-2) for family dissolution. The two facets work together rather than duplicating each other.
Access the full noslegal taxonomy (all facets, extension packs and process maps) at noslegal.org.