Reimagine the access to technical documentation with the support of artificial intelligence

Many manufacturers own miscellaneous inventory product documentation in need of supply to their customers and users. For an efficient retrieval in a content delivery system (CDS), these documents should be first labeled with respect to classification standards relevant for a given industry. However, due to the lack of resources in Technical Editing departments, this classification is often missing, incomplete or follows non-standardized procedures and, as a result, valuable information in the product documentation cannot be discovered in a CDS.

In this project, we aim to approach the problem of large-scale access to product documentation by leveraging the methods of artificial intelligence (AI) for automated document classification. Our goal is to automatically generate metadata for ca. 100’000 documents and make them easily accessible via digital media. We are looking for industry partners seeking to make the information in the company’s product documentation discoverable, and willing to share product documentation with the research partner for experiments. The industry partner(s) will benefit by having its documentation labeled and ready for use in a CDS, and by increasing its digital offer with the developed document discovery tool. Furthermore, we are looking for both complementary industry partners – service or software sector – and research partners to improve applied knowledge. These complementary industry partners might benefit from scaling the solution and providing it to the market.

Conventions of Innosuisse funding projects are the participant concept, not only in terms of content, but also on the administrative-financial level in return gaining the scientific knowledge and expertise of one or several research partners. The contribution of potential industry partner(s) consists of providing non-classified inventory product documentation, being interested in the application or implementation of a Content Delivery Portal and sharing entrepreneurial know how. Financially, the project costs are partially incurring within your company and shall be covering at least same the amount as the funding offered by Innosuisse, of which at least 10% shall cover your research partner’s expenses in terms of a cash contribution.

The planned project intends to be an application-based study where research institutions will work together with at least one industry partner. During this Innosuisse project, we will conduct a qualitative and quantitative text analysis of inventory product documentation. The purpose of this project is to identify the most effective way to automatically produce metadata within a certain classification standard (e.g. iiRDS or VDI 2770). We are also planning to develop a chatbot as a user- friendly interface for document search in a CDS, translating queries in natural language into machine- readable search requests, allowing the future users to interact with the search engine in a comfortable manner.

Economical value for the industry partner (s) is estimated in securing a competitive standing as innovative company e.g. concerning user friendliness or linguistically simplified access resulting in potential revenue and reputation increase. On one hand, the aspired solution is intended to be applicable with the existing resources in Technical Editing, on the other hand, the digital offer might potentially increase customer acquisition, customer satisfaction as well as customer retention. A social value is anticipated for users who could directly profit from the digital supply in terms of optimized language access. An overall economic, social benefit might be gained by actively promoting the automated processes by digitalization in technical communication. As a result of this digital shift, actual challenges like the increasing complexity in the working environment accompanied by the growing skills demand or the prospective labor shortage might be partly resolved.