Generative AI + Customer Service - Use Cases

[former_member191690]

With recent advances in Generative AI, large language model use cases have really come to life!  It's not just interpreting natural language but also understanding semantic meaning behind across large corpus of language text and generating new content.  This is indeed next phase of applied AI.

As I played with it, here are few use cases I find compelling -

For Agents -

• Generate Agent response - auto compose the response - every customer interaction, eg email - Automatically composes email draft, leveraging Generative AI . Agent to just review and hit send. After a while this could be fully automatized.  .


• Recommend course of action - Next Best actions -   For every customer inquiry - suggests possible course of actions.  For varied support requests - Order Support,  Returns, Product Troubleshooting, Invoice enquiry… This could be based on the knowledge spread across various sources - service bulletins, alerts, notifications, help guide,  SharePoint documents.


• Generate Visual flow chart of the problem -  From the long back-forth communication. Map the text to images and generate visual flow.  Make it easier for new Agents to understand the problem.

For Customers -

• Self Service - More complex queries -  what is my unit cost with bulk orders ? Does my contract allows me to do this ?  Delivery delays - Am I the only one .. Are my competitors  facing same issue…?  These would be based on training with combination of public and private data.

For Service Manager -

• More deeper insights from analyzing call logs, voice transcripts, CSAT scores, service procedures, resolution history .. etc


• Queries include -  1. How can I improve resolution times…. 2. How can I improve Agent efficiencies without compromising CSAT. 3. Best practices - what are the best practices for logistics handling

For Consultant -

• This is most popular, I can get code snippets to do my job. Eg. - Can you show be code for creating custom actions?  See response (screen shot)

ChatGPT response

These would be more impactful …..if we find a way to make it more domain/customer specific by leveraging industry/customer data for training (with permissions)

These are my initial thoughts. Just scratching the surface.  There would be lot more use cases we all can think of.  Exciting times ahead.

Disclaimer:
SAP notes that posts about potential uses of generative AI and large language models are merely the individual poster's ideas and opinions, and do not represent SAP's official position or future development roadmap. SAP has no legal obligation or other commitment to pursue any course of business, or develop or release any functionality, mentioned in any post or related content on this website.