5.5 Document preview

The document preview in INTERGATOR Patent Search provides a complete, interactive view of an individual patent result. It enables in-depth detailed analyses, the comparison of drawings, the quick creation of new filters directly from the text, and the start of publication-related AI analyses.

The document preview is therefore not only a reading view, but also a central starting point for the technical evaluation of individual patent documents. From the preview, technical content can be analyzed, key statements can be summarized, relevant text passages can be checked and a publication report can be created for the opened patent document.

The preview opens when clicking on a result, e.g. via the context menu → “Preview”, and consists of three clearly structured areas:

• Bibliographic information: title, publication number(s), status, applicant, inventor, IPC class, technology domain
• Automatically structured text sections: abstract, claims, description
• Search term highlighting: relevant terms from the search query are highlighted in color
• Directly linked filters: clicking on, for example, applicant, inventor or IPC class automatically sets a corresponding filter

The left column supports quick orientation within the patent document. The colored highlights make it possible to identify text references related to the current search directly. Bibliographic information and linked elements can also be used as a starting point for additional filters and follow-up searches.

• Technical drawings: central figures of the patent with a preview bar (thumbnails) for navigation
• AI tasks: dropdown menu with predefined prompts for publication-related analyses, e.g.:
  1. Free question / instruction
  2. Validity analysis
  3. Relevance re-ranker
  4. Compare claims
  5. Parameters with values and units
  6. Claims (brief technical summary)
  7. Claim features
  8. Claim features (expanded from description)
  9. Claim features with synonyms
  10. Brief description
  11. Innovation
  12. Component graph
  13. Title and abstract in German
  14. Brief description in German
• Patent family: list of all associated patent documents
• Cited patent literature: references to cited and citing documents
• Similar documents: AI-based calculated thematic similarities

The right column combines visual document analysis with AI-supported evaluation. The technical drawings help to understand components, processes and functional relationships more quickly. The AI tasks can be applied directly to the opened patent document and provide structured results, such as technical summaries, claim analyses, feature extractions, parameter lists, innovation assessments, component graphs or translations.

Some AI tasks refer exclusively to the currently opened publication. Other functions are comparative and require an additional reference context, such as a source or reference document from an assistant workflow or an active semantic document filter. This applies in particular to functions such as validity analysis, relevance re-ranker and compare claims.

• Display of all currently active filters, e.g. corpus, technology domain, IPC
• Option to deactivate or remove individual filters with one click
• Progress indicator for loading additional results

The interactive bar shows the current search context and makes it possible to quickly check, deactivate or remove set filters. This allows the search to be further refined directly from the document preview.

The document preview is the central place for publication-related AI analyses. The AI tasks each refer to the currently opened patent document and support the quick technical classification of the document.

Typical use cases include:

• free questions or instructions relating to the entire document or specifically to individual full-text fields such as title, abstract, claims or description,
• technical summary of a patent document,
• analysis and summary of the claims for a better technical understanding,
• extraction of claim features from the claims,
• expansion of claim features using additional information from the description,
• derivation of synonyms for claim features for further searches,
• extraction of parameters including value ranges and units,
• assessment of the innovation based on the description and in comparison with the prior art cited in the document,
• visualization of components and features in a component graph,
• translation of title, abstract or brief description into German,
• as well as comparative analyses with reference to a reference document.

The results of the AI analysis are displayed directly in the context of the preview. This allows the user to compare the generated statements with the original text passages, drawings and bibliographic information. The AI analysis does not replace technical review, but supports faster understanding and the structured evaluation of the patent document.

Some AI tasks in the document preview do not only analyze the currently opened patent document, but compare it with a reference context. These include in particular:

• Validity analysis – analyzes the claims of the reference document in the query with regard to possible challengeability by the currently opened publication.
• Relevance re-ranker – analyzes the claims of the search result with regard to their relevance in relation to the reference document.
• Compare claims – compares the claims of the opened publication with those of the reference document in the query.

A reference document is typically created through an assistant-supported workflow, for example by entering a publication number in the validity search, FTO search or monitoring search. The corresponding patent document then serves as the source or reference document of the generated query.

If no suitable reference context is available, comparative AI tasks cannot be executed meaningfully or will not provide a reliable assessment. In this case, a suitable search assistant should first be started, a source document should be defined or a semantic document filter should be set.

The publication report always refers to a single patent document. It is used to prepare the most important information, technical statements and analysis results for a specific document in a traceable way.

A publication report can be started directly from the document preview. The starting point is the currently opened patent document. The report summarizes the publication-related information in a structured way and can be used as the basis for internal evaluations, technical analyses, discussions or further search decisions.

A publication report may include, among other things, the following content:

• bibliographic information on the patent document,
• brief technical description,
• key features and components,
• relevant text passages and references,
• technical functional relationships,
• summary of the claims,
• information on patent family, citations or similar documents,
• as well as a documentable assessment of technical relevance.

It is important to distinguish this from the search report: the publication report refers to a single patent document. A search report, by contrast, typically documents a search, a result list, several relevant documents or an overarching analysis.

In the document preview, relevant text passages can be reviewed directly in the patent document and used for further search. Marked terms, sentences or sections can serve as the basis for new semantic filters. This allows the search to be refined in a targeted way directly from the document.

This is particularly helpful when:

• a specific technical feature is to be followed up,
• a section is particularly relevant to the question,
• the current result list is to be technically refined,
• or a document is to serve as the starting point for similar documents.

By combining original text, highlighting, drawings, AI tasks and filter functions, the document preview enables a particularly close connection between reading, understanding, analyzing and continuing the search.

The document preview therefore provides quick, structured access to all contents of a patent – ideal for detailed analyses, comparing technical content, publication-related AI analyses, publication reports and targeted follow-up work with relevant filters without leaving the results page.