What good taxonomy implementation looks like across delivery, knowledge, pipeline and people — and how value compounds when the same concepts are used consistently across all four.
This chapter addresses what good taxonomy implementation looks like across the four areas of need introduced in chapter 2. The value compounds when the same concepts are used consistently across all four, and the guidance below should be read with that in mind.
Delivery covers the scoping, planning, pricing, execution and management of legal work. Accurate matter classification is the foundation on which the other three areas depend: it is what makes knowledge retrievable, credentials credible and experience data meaningful.
The guidance below sets out recommended practice for applying each relevant facet to matter classification. For legal technology companies, it describes the classification model that products supporting delivery should be designed around.
The implementation points in section 4.1 apply in full. The key delivery-specific point is the treatment of matters with mixed work types. Where a matter develops a substantial element of a materially different second-level work type (for example, where a corporate transaction gives rise to significant litigation), a new sub-matter or linked matter should be opened for that work. Mixing materially different work types in a single matter undermines financial reporting, interferes with like-for-like comparison and muddies the classification context available for knowledge management and AI. Compliance will not be perfect, and a sensible materiality threshold should be defined, but a practically useful level of compliance can be achieved with clear guidance and periodic review.
The implementation points in section 4.2 apply in full. Unlike work type, a "one and only one" requirement at the second level may be artificial: multiple areas of law often apply to a single matter in complex and entangled ways that cannot readily be separated into sub-matters. One practical approach is to require a single primary area of law while allowing additional areas to be tagged at any level. Third-level topics are best treated as optional, used for knowledge and experience purposes without the rigour required at the second level.
The key role to capture in all matters is the one connected with the relevant work type (for example, defendant in litigation, buyer in a transaction, borrower in a financing). A one-and-only-one rule should apply to each matter or sub-matter for this primary role.
It is also worth capturing a role connected with the area of law (for example, employer for employment law matters). This adds a useful dimension for financial analysis, experience profiling and knowledge retrieval.
In implementation, the availability of roles in the software should be linked to the work type or area of law already selected, showing only roles relevant to dispute resolution when the work type is dispute resolution, for example. This reduces the risk of irrelevant roles being selected.
The implementation points in section 4.4 apply in full. The key delivery-specific point is the distinction between client sector and matter sector. If a pharmaceuticals company acquires an office building, the matter is more usefully classified in the real estate sector than the pharmaceuticals sector, because the relevant experience and knowledge concerns real estate rather than the client's industry. The client's sector is a separate data point, captured on the client record.
Information types, uses, status and audience are intended to be used together to classify usefully the things found in legal service provider or legal department messages, document repositories and other records.
The grid in the taxonomy spreadsheet illustrates how type and use may be combined to provide a rich classification.
We suggest that implementation should enforce
The implementation points in section 4.5 apply in full, including the guidance on distinguishing governing law, dispute resolution forum and asset or event location. Where the work is being done and where the client is based are typically captured elsewhere in the matter management or financial system and do not need to be duplicated here. A one-or-more approach will often be appropriate given the cross-border complexities that can arise.
Process elements are different from the above facets in that they contain process maps rather than concepts for classifying matters. They are most valuable when applied as part of a systematic approach to process improvement, project management and pricing. Even without such a systematic approach, applying process elements to time narrative data can help identify financial and other impacts requiring attention.
Knowledge, as defined by ISO 30401, is a human or organisational asset enabling effective decisions and action in context. In legal work, that context is precisely what the noslegal facets describe: the type of work being done, the area of law involved, the role of the client, the sector and the place. Classifying knowledge against those facets is therefore not a filing exercise. It is what makes knowledge findable at the moment it is actually needed.
The shape of a knowledge base differs between legal services providers and in-house teams. A law firm's knowledge typically organises around practice areas and work types. An in-house team's knowledge tends to cluster around the recurring legal issues of the business it supports, the policies and standards it maintains and the relationships it manages with external providers. The guidance below applies to both, with differences noted where they are significant.
Not all legal knowledge should be treated the same way. Different types arise in different ways and benefit from different approaches to capture, classification and governance.
Much legal knowledge is created in the course of handling matters: advice given, drafts and mark-ups, emails explaining legal reasoning, supporting materials. The most important requirement for this material is that it remains associated with the matter in which it arose. If the matter has been classified using the noslegal facets, the associated work product automatically inherits that context. In practice this means storing documents within the relevant matter workspace, ensuring the matter itself is classified accurately, and avoiding attempts to classify every individual document separately.
Some materials generated during a matter capture particularly important steps or decisions. These can be thought of as key matter records: the curated spine of a matter worth preserving and making findable beyond the immediate context in which it was created.
Over time, some materials prove useful beyond the matter in which they originated and become knowledge assets (precedent clauses, example advice, guidance notes and similar). Knowledge assets may also be created deliberately outside any particular matter: training materials, policy documents, articles and other publications are typical examples.
For knowledge assets, deliberate classification becomes more valuable than inherited matter context. Assets can typically be tagged using selected noslegal facets (work type, area of law, place and sector being the most commonly useful) along with any organisation-specific fields such as practice group or business unit. A small number of consistently applied fields is almost always more valuable than a larger number applied inconsistently.
A small subset of knowledge assets may become particularly important to how an organisation performs legal work: key playbooks, canonical templates, structured guidance on recurring issues. Because of their importance, these assets often benefit from stronger governance, including clear ownership, alignment with authoritative sources such as law or internal policy, and periodic review to ensure they remain current. In an in-house team, this category often includes standard contract templates, approval frameworks and legal policies that the whole organisation relies on, making governance especially important since the consequences of outdated material can be significant.
Legal knowledge often develops through stages: everyday work product created during a matter; important materials recognised as key matter records; selected materials curated as knowledge assets; and a small subset maintained as high-value knowledge assets with stronger governance. Taxonomy plays a different role at each stage. Early stages rely mainly on matter classification, with knowledge inheriting its context from the matter. Later stages benefit from deliberate tagging and active governance. Understanding this lifecycle helps organisations focus classification effort where it adds most value rather than attempting to classify everything equally.
The noslegal taxonomy can also be used to describe the experience of individuals within an organisation. When matters are classified consistently, the experience of the lawyers who handled them can be aggregated and analysed. Individuals may be associated with the work types they handle, the areas of law in which they have experience, the sectors in which they work, the jurisdictions in which they operate and the roles they typically perform. Using the same taxonomy across matters, knowledge and people makes it easier to identify relevant experience when new matters arise. This topic is developed further in section 6.4.
Effective classification using noslegal has a particularly direct payoff in AI-assisted knowledge retrieval. Where a language model draws on a corpus to answer questions or surface relevant materials, this will perform better if knowledge assets are consistently tagged by work type, area of law and so on. Those classifications can be used to filter the corpus before the model processes it. Without this pre-filtering, the model is working across a larger, noisier set of materials, with implications for performance. An investment in consistent classification is therefore an investment in effective use of AI.
Pipeline covers the activities by which legal organisations understand, anticipate and influence what work is upcoming or available. For legal services providers this includes marketing, business development, credentials and client relationship management. For in-house teams it includes forward planning, resourcing and managing the flow of work from the business.
The most valuable pipeline data is largely the same data generated for delivery. A well-classified matter record (capturing work type, area of law, sector, place and role) is also the foundation for credentials, experience profiles and portfolio analysis. An organisation that classifies its matters accurately for delivery purposes gets most of the pipeline benefit without any additional classification effort. The primary pipeline requirement is therefore not a separate classification exercise but a commitment to delivery classification quality, combined with the ability to surface and aggregate that data in the formats pipeline activities require (pitch documents, capability statements, client reporting, demand analysis and so on).
Where separate pipeline data exists (in CRM software, a credentials database or a business development platform) the key requirement is that it is synchronised with the matter and financial data generated in delivery. A credentials record that describes a matter differently from the way it is classified in the matter management system undermines the credibility of the credential and makes portfolio analysis unreliable. Synchronisation does not necessarily mean identical fields, but it does mean that the core taxonomy classifications applied to a matter in delivery are reflected consistently wherever that matter appears in pipeline systems.
Some pipeline activities generate data with no direct equivalent in delivery (information about prospective clients, pitch outcomes, or the status of business development relationships). Where pipeline records are linked to matters (for example, where a pitch record is linked to the matter that resulted from it) applying consistent taxonomy classifications to both makes it possible to analyse conversion rates, pitch success and demand patterns by work type, sector or geography.
The people area covers understanding and communicating what experience and skills an organisation's legal professionals have, identifying gaps, and making good decisions about work allocation, recruitment, training, secondments and career development. For legal services providers it includes the credentials and experience profiling that supports business development. For in-house teams it includes managing the balance between work done internally and work sent to external providers.
As with pipeline, the most valuable source of people-relevant data is that generated in delivery. A lawyer's experience is primarily demonstrated by the matters they have worked on. If those matters are classified consistently and accurately (by work type, area of law, sector, place and role) a meaningful and up-to-date picture of each individual's experience accumulates automatically as a by-product of ordinary matter management. This is both more reliable and more granular than experience data that depends on individuals self-reporting their expertise, updating their profile or completing periodic HR exercises that are easy to neglect.
Separate HR or marketing databases serve important purposes (recording formal qualifications, training history, performance data and curated narrative profiles used in pitches and on websites). But these sources tend to become outdated, to reflect how individuals wish to be seen rather than what they have actually done, and to be maintained inconsistently. Matter-derived experience data provides a more objective and continuously updated foundation that separate systems can draw on and supplement rather than substitute for.
The noslegal facets most relevant to experience profiling are work type, area of law, sector, place and role. Process elements data can also be relevant in some cases (for example, to identify that someone has a particular focus on ediscovery rather than other aspects of litigation). When matters are classified consistently against these facets and individuals are associated with the matters they have worked on (whether as supervising partner, lead associate or in another defined capacity) it becomes possible to aggregate and analyse experience across the organisation in a structured way.
Useful questions that well-classified matter data can answer include: which lawyers have handled matters of a given work type and area of law in a particular sector or jurisdiction; who has experience of a particular role (acting for defendants in arbitration, or advising borrowers in leveraged finance transactions, for example); how broadly or narrowly experience is distributed across the team; and where gaps exist relative to the work the organisation is doing or anticipates doing. This data has practical value across several contexts: identifying the right person for a new matter, supporting supervision and development conversations, informing recruitment decisions and demonstrating credentials in pitches and capability reviews.
Where HR systems, learning management platforms or marketing profile databases exist alongside matter management systems, the same synchronisation principle applies as in pipeline. The core taxonomy classifications from delivery should be reflected consistently wherever individual experience is described. A lawyer's profile on the firm's website or in a pitch document should be grounded in, and consistent with, their matter history as classified in the matter management system, rather than maintained as a separate and potentially divergent record. Achieving this in practice often requires deliberate integration work or periodic reconciliation processes, but the direction of travel should be clear: matter-derived data is the authoritative source for what people have actually done, and other systems should draw on it rather than operate independently of it.
Experience data derived from matter classification tells you what people have done. It does not by itself tell you what they are capable of doing, how well they have done it, or what they need to do next to develop. These questions require additional data (formal skills assessments, supervision feedback, training records, development plans) that sits largely outside the scope of noslegal as currently constituted.