Technology sector and subtechnology atlas

Technologies: evidence-led reference

This atlas uses human-readable technology families as a gateway into patent-information research. An industry term such as 'artificial intelligence' or 'solid-state batteries' is not itself an official IPC or CPC class. The project's two source taxonomies have 50 and 54 sector headings, respectively, with differences in grouping. The intended product is a documented many-to-many crosswalk that preserves each source taxonomy and adds technical-function evidence before suggesting official classifications. No patent corpus, ownership ranking or actual classification mapping should be asserted solely because a sector name appears in the planning register.

Sector, family and subtechnology are different levels

A sector is an editorial organizing unit, such as robotics or biotechnology. A family narrows the technical function, such as robot manipulation or protein engineering, and a subtechnology describes a more particular mechanism or task. A single invention may combine several families and multiple sectors, so the hierarchy cannot be exclusive by default. The atlas should store stable identifiers for each node, a human-readable title, alternative names, narrower and broader relationships, and the question a reader is trying to investigate. Where an industry classification changes with marketing fashion, retain the original source term and create an explicit alias rather than rewriting every related URL.

The 50-sector and 54-sector registers are source records

The September portfolio blueprint proposes a 50-sector technology ontology intended for four sites with different search intents. The later legal-services research uses 54 categories organized for professional practice and market-intelligence questions. This is not proof that 104 independent sectors exist. Reconcile names through explicit crosswalk relations: equivalent, narrower, broader, overlapping, convergent or unresolved. A relationship such as 'medical AI overlaps digital health and machine learning' can be recorded without collapsing the legal, clinical and computing dimensions. Each crosswalk decision should include rationale, date, reviewer and consequences for existing canonical URLs.

Primary reference: WIPO IPC reference. Interpret only within the source’s published jurisdiction and scope.

Map technical functions before patent symbols

A reliable classification lookup begins with the invention's mechanism, not its commercial label. For example, an AI-based imaging system may involve data processing, image acquisition, diagnosis and hardware components. The method should identify those functions independently, search the official IPC and CPC schemes, read inclusion notes and confirm representative patent publications. A broad technology page can display candidate areas and research routes; exact subclass or group claims require source-level verification. Many-to-many mappings are normal because official classes are organized around technical characteristics that do not correspond one-for-one with business industries.

Handle convergence as relationships

A technology such as quantum-enhanced sensors or robotics using spatial computing can live in multiple parent branches. Creating duplicate child articles under every possible sector creates competing canonicals. Instead record a primary canonical technology node and cross-reference neighboring sectors by stable identifiers. A convergence page needs an additional technical question: what mechanism or problem emerges from the combination, beyond summaries of two existing sectors? A cross-sector node is justified only when the corpus or method differs materially. Where the distinction is not evidenced, use an explanatory paragraph and links rather than another indexable page.

Primary reference: Official CPC scheme and definitions. Interpret only within the source’s published jurisdiction and scope.

Use alternate terminology carefully

Search queries may use commercial names, laboratory terms, patent examiner language and acronyms for the same or partially overlapping technique. The atlas can document these without suggesting they are exact synonyms. Agentic AI and autonomous software agents, for example, may overlap yet apply differently in a patent document. Store an alias type: lexical synonym, related concept, historical name or ambiguous shorthand. Technical phrases should be linked to definitions and examples. When translating terminology across languages, keep original language text where practical and avoid projecting an English term into a jurisdiction whose official classification notes are different.

Use a technology-node evidence record

A useful node should provide the original definition source, illustrative technical problem, possible solution families, adjacent subjects, classification candidate links, source edition and research uncertainty. The next analytical layer may include owners, trends and patent counts, but only after a defined corpus has been collected and deduplicated. This .org atlas is not that completed market analysis. The separate research property can evaluate technology-specific landscapes while this site explains how a reproducible question and technology taxonomy are constructed. Include references to authoritative patent classifications and maintain a glossary for overlapping terms.

Publish only distinct subtechnology references

A separate page is justified when it provides a different technical-function explanation, source mapping or research method from the parent. A new slug should not exist simply because one industry term has many brand synonyms. Candidate subtechnology records can remain in the private registry until the evidence packet establishes a useful independent canonical. The reader should be able to move from sector to family, from family to method, and from method to official source without encountering vacant pages. A published hub is a documented conceptual map, not a list of empty routes or an asserted count of opportunities.

Primary reference: WIPO patent information. Interpret only within the source’s published jurisdiction and scope.

Illustrative research scenario: applying the distinctions

Take the term 'agentic AI' as an illustrative node. It may describe systems in which software components plan and coordinate tool actions, but marketing descriptions vary. Under a source-led taxonomy, the sector might be AI, the family autonomous software agents and the subtechnology a particular tool-routing or safety mechanism. Another node for 'legal AI' describes an application domain, not necessarily a new technical family. Both may connect to document retrieval or access-control technologies. A useful atlas records these as broader/narrower or overlapping relationships and explains the mechanism rather than forcing a one-to-one class correspondence. To suggest official patent classifications, a researcher would inspect the actual invention features, then verify candidate IPC/CPC entries and relevant published examples. This distinction avoids creating multiple near-identical pages for 'AI agent', 'agentic software' and 'autonomous legal assistant'. The page may teach the node structure while deferring unverified market counts or definitive class assignments.

Practical decision and evidence matrix

What this reference establishes — and what it does not

The material above is a source-bounded reference concerning technologies. It distinguishes official documentation from editorial method and provides practical decision boundaries; it does not claim an exhaustive global search, an empirical dataset that was not collected, a legal opinion, or certified implementation of publisher policy. Readers seeking to apply a source to a current filing, claim or professional obligation should inspect the applicable original authority, record its effective date, and obtain the necessary qualified advice. A page is updated substantively only when its evidence, method or answer materially changes.