LLM’s can combine with voice recognition to enable interaction with existing enterprise software. Our use case example would involve a user communicating with the LLM through voice recognition, where the spoken prompt activates existing user interface actions on the enterprise software. The envisaged use case is that simple user interfaces of enterprise software can be rendered on handheld devices with navigation and actions being prompted by voice. The question is how consistent is the existing navigation of enterprise systems? Since a key requirement is the consistency of the existing navigation, so that the spoken prompts used to involve the LLMs can be sufficiently consistent. 
Additionally, we are testing how LLMs can be used to assist in the organization of enterprise emails. The incoming email stream would be parsed by the LLM and assigned to enterprise processes of interest to the stakeholding of different enterprise users. Both the e-mails and their attachments can be recognised as belonging to processes and assigned to them. The value here is that the daunting inboxes of emails and incoming documents that every enterprise is familiar with, can be organised according to fine-grained processes within the enterprise. So that enterprise users are assisted by having the system organise and instantiate the incoming stream of emails and documents instead of having to read and organise the incoming stream unassisted, and while enterprise systems remain unconnected to the stream of content.