Since last November, everyone has been talking about ChatGPT and thinking about what tasks AI can do for us.
I believe that AI has great potential when applied to social media interaction. There is no doubt that product blogging on the SAP Community is an effective way to provide customers and users with broader and more detailed how-to information, but that is still far from enough. Let’s take one of my blogs as an example to show how we can use AI to boost the range of our blogs on SAP Community.
What I do after publishing blogs
I published a how-to blog on April 28, 2022, explaining how to configure a journal entry workflow. In the blog, there are some examples of complicated scenarios: How to Create Workflows that Require Multilevel Approval. We were very pleased to see that it received 5000+ visits in the past year, and quite a few people, most of whom are consultants, have left questions asking for specific solutions to suit their customer’s business cases. As you can see from the comments section, some people asked how to use certain BAdIs, which is relevant to this blog, but only slightly. In fact, many of these questions are not easy to answer, even for our development experts, but we still need to try to provide the best response we can - because customers are waiting for answers, and blogging is the only channel where they can talk directly to development teams, other than creating tickets. As the contact for blog questions, I usually go through this simple process before replying to users and customers:
I forward the question on the blog page to the developer expert responsible for the topic.
I schedule a short meeting with the developer to identify what the question is about. • If the question is clear and easy, we reply with a quick solution as a comment on the blog page. • If further explanation is needed, we start an email communication or a phone call with the user.
The developer begins to research and test. This usually takes some time, and sometimes the knowledge is a little beyond the developer’s frame of reference. In this case, we need to check with colleagues in other roles.
We consolidate our research and test results and provide the solution.
Some blog questions don’t require this process. I am very glad when I see that existing documentation, say, on the SAP Help Portal, can provide the answer. In such a case, I can quickly find the links to the product assistance and forward them. Unfortunately, few questions can be addressed in this way; most questions need a consolidation of knowledge and sometimes implementation and testing. Honestly, I wasn’t expecting all this follow-up work when I started blogging! I believe that someday AI will become the knowledge expert to help with this work, and my job will be to verify that the content AI collects is correct, before sending it to customers. That would be perfect, both for us and our users.
How can AI help? Based on my general AI understanding, it is not difficult now to train AI with knowledge tailored to your specific requirements. As a technical writer who needs time to create blogs and respond to the questions they bring, I hope AI can help with efforts spent on tasks such as collecting relevant information, analyzing it, and coming up with an initial solution, whether in text or graphics.
Will all of these wishes come true? I believe the moment is just around the corner.
SAP has built an enormous (and growing) knowledge base, both externally and internally. The resources come from various deliverables on the SAP Help Portal, KBAs, wiki pages, design docs, specs, blogs on communities, and other sources. These documentation sets are managed by many different people and organizations, and it is unlikely that one person will know where to find the answer to every question among all the materials available. However, this is an achievable task for AI. After we feed AI the knowledge that it needs to know, it can easily suggest a possible solution, depending on the model chosen and how it is trained. Of course, for the sake of information safety, we as human beings must carefully verify the content before exposing the information to the outside.
Looking ahead, I can think of the following changes AI might bring to social media:
Information centralization and consumption: Not only can AI access and combine all information, but it can also process the content and the connection as a human brain does.
Personalization and Customization: AI can provide very specific solutions, much like a consultant who fully understands a customer’s scenarios or issues, if specific-enough questions are asked.
Company policy compliance: When communicating with customers, AI can be trained to be compliant with company policies and social media guidelines.
Multiple language support by default: AI is a born language expert, and it can generate output in multiple languages without a translation.
It looks very promising for AI to join in our social media work, but we still need to be very careful to monitor AI behavior and output. New requirements are waiting for us: to communicate with AI using an engineer’s process while having a scientist’s mind. It’s always challenging to deal with technological change, but these challenges create the momentum that drives progress.
Are you ready to work with AI for social media?
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.