US 2026/0093716 A1Published application

Systems and methods for clustering cloud service requests

AssigneeSap SE

InventorsRobert LOWSKI, Florian FRICKE, Varsha SHETTY

Life of this patent

Estimating the term…

Loading patent…

Abstract

Described herein are techniques for analyzing assisted service requests received so that the service provider can identify new services to create for purposes of enriching the service catalog. The assisted service requests can be analyzed and organized into a plurality of clusters based on similarity. The service provider may review the clusters when deciding which new service requests to create for the service catalog. In one example, clusters containing more assisted service requests may be prioritized over clusters containing fewer assisted service requests since the newly created service request would be able to service a larger number of customer requests.

Claims20

Dependent claims

A method comprising:

retrieving a service catalog, the service catalog comprising at least one cloud service and one or more service request templates associated with the at least one cloud service;
receiving a plurality of service requests that are based on the one or more service request templates;
receiving a plurality of assisted service requests that are not based on the one or more service request templates, wherein each of the plurality of assisted service requests comprise a text description in natural language, wherein the text description describes a request for a cloud service that is not in the service catalog;
generating a large language model (LLM) embedding for each of the plurality of assisted service requests;
generating a plurality of clusters based on the LLM embedding for each of the plurality of assisted service requests, wherein each of the plurality of clusters comprise one or more assisted service requests from the plurality of assisted service requests;
applying a generative artificial intelligence (AI) model to the plurality of clusters to generate metadata for each cluster in the plurality of clusters;
and
generating a graphical user interface (GUI) comprising the plurality of clusters, the metadata for each cluster, and the one or more assisted service requests of each cluster.

5 more elements

A non-transitory computer-readable medium storing a program executable by one or more processors, the program comprising sets of instructions for:

retrieving a service catalog, the service catalog comprising at least one cloud service and one or more service request templates associated with the at least one cloud service;
receiving a plurality of service requests that are based on the one or more service request templates;
receiving a plurality of assisted service requests that are not based on the one or more service request templates, wherein each of the plurality of assisted service requests comprise a text description in natural language, wherein the text description describes a request for a cloud service that is not in the service catalog;
generating a large language model (LLM) embedding for each of the plurality of assisted service requests;
generating a plurality of clusters based on the LLM embedding for each of the plurality of assisted service requests, wherein each of the plurality of clusters comprise one or more assisted service requests from the plurality of assisted service requests;
applying a generative artificial intelligence (AI) model to the plurality of clusters to generate metadata for each cluster in the plurality of clusters;
and
generating a graphical user interface (GUI) comprising the plurality of clusters, the metadata for each cluster, and the one or more assisted service requests of each cluster.

5 more elements

Dependencies are inferred from the claim text.