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.

[Abhijit_Shinde]

Great start Kiran, Thanks! This is a food for thought for everyone to incorporate more of AI in day-to-day tasks and switch efforts to value added ones.

[sanchitahirrao]

Very interesting. Thanks for sharing these, Kiran! I played around with the SAP version of Chat GPT (SAP Enterprise Playground GPT-3 (ondemand.com) this week and I was very intrigued by its immense possibilities. Have brainstormed a few use cases here for Service Cloud.

• Knowledge Bases
  • Based on empty search results for KB articles, our AI service can recommend the admin to come up with new articles on those missing topics and when creating a new article, we can leverage Chat GPT to write those missing content for us. (at-least 80% of it). This makes creating new KB articles very easy and the recommendations is provided by our AI Service.
• Walkthrough or Help Content on On-boarding
  • I noticed that in Service Cloud, the in-built product help section itself can be improved by leaps and bounds using Chat GPT. This will reduce the TTV for the new customers who are on-boarding to Service Cloud and are trying to figure how to get things done. Ex: When I ask  ‘How to create a new Org Unit’ in the help section it gives me no results found. But, using Chat GPT we can provide an ‘Help Bot’ to help users with their daily tasks with not just ‘How to create a new Org Unit’ but ‘Create a new Org Unit for me with the name XYZ, etc.’

• Service Manager Use Cases
  • Workforce Performance Analysis – Our AI service can compare the performance of one or more agents or Service Units based on certain metrics like CSAT scores, resolution history, etc. and rank them quantitively based on these parameters. Chat GPT can them provide perspective next steps on what the mid to low ranked agents are lacking on which variables and prepare a feedback memo to send to them.
  • Few prompts a Service Manager can provide to Chat GPT are as follows,
    • Information Collection:
      • “Give me a list of agents in the Org Unit ‘XYZ’ who have expired Skill Assignments.”
      • “Give me a list of employee with the Business Role ‘Service Agent Canada’ who are having the Skill ‘French’ with proficiency A2 and above.”
      • “Which agents who are reporting to me have the lowest call resolution rates.”
      • “Compare the CSAT of Org Unit 1 and Org Unit 2 over the past 3 months.”
    • Action Performing:
      • “Assign all the agents with Skills as French to the Org Unit ‘Canada Service’.”
• Admin Use Cases
  • As we know an Admin’s main responsibility is to 1.) Setup the various entities in Service Cloud 2.) On-board people to it 3.) later go on to perform daily routing maintenance tasks.  We can make their life easier with an ‘Co-pilot’ mode using a combination of our AI Services and Chat GPT.
  • This ‘Co-pilot’ will be an admin only assistance tool, which will help Admin with all the 3 above responsibilities.
    • Setting Up: When an admin enters Service Cloud for the 1^st time they have a daunting task of creating multiple entities like Org Units, Employees, importing data, etc. Chat GPT can provide prescriptive step by step instructions on the things to do next in a sequential manner. It will be sort of like an ‘Admin Setup Wizard’.
    • On-boarding people: It will provide recommendations on how to import the employees based on the existing data base, provide recommendations to group them under Org Units and assign them to their required Business Roles, etc.
    • Maintenance & Daily Tasks: There is so much value and automation which can be done here. Ex: When a new employee joins the org, the on-boarding can be automated if not at-least the next steps can be provided to them admin when they say to their ‘Co-pilot’ that “Hi! A new Service Agent, John is joining our organization with email ID john@abc.com”. The AI service and Chat GPT will provide the next steps with ‘action buttons’ for the admin to proceed and certain details will be pre-filled with recommend ‘Business Role’ based on the existing agents.
      • Data Clean-up & Redundancy: Over a period of time, there might be false, incomplete, redundant data in the system like an important filed missing for a Key Account or a missing location for a Registered Product, etc. Admin can tell Chat GPT to ‘Initiate data clean-up’ and with the help of our AI services it can look at the entire system and provide recommendations on what needs to be done.

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.