Salesforce ChatGPT and the Salesforce Ecosystem
ChatGPT is an AI chatbot or Large Language Model developed by OpenAI, with backing from Microsoft. It uses NLP to interact with users, providing answers to questions by scanning its vast database of information obtained from millions of websites. ChatGPT presents the answers in a natural, conversational style, though accuracy may vary based on the specificity of the request.
The Salesforce ecosystem could potentially be impacted by ChatGPT, with both winners and losers. Its best use cases include creating code such as validation rules, formulas, and Apex classes, but it should not be relied upon to generate declarative results or provide accurate answers to broader scopes. Junior developers may be at risk, as ChatGPT has learned from millions of lines of code and could outperform them in writing code. It is also fair to note that Salesforce has utilized artificial intelligence since 2017.
Business analysts and consultants may find ChatGPT helpful in generating generic proposals or suggesting process steps, but should be wary of relying too heavily on its results. Architects, on the other hand, may not find much value in ChatGPT as it lacks the experience and context necessary for the role. ChatGPT may also be useful for generating first drafts of documents like training materials or business cases, but the user must specify the reasoning behind changes, which ChatGPT is incapable of knowing.
ChatGPT is essentially a content development tool, not to be confused with Salesforce Service Cloud Voice and chatbots. Read more about them here. If you currently utilize chatbots be aware of Salesforce’s My Domain changes and how they can impact them.
For marketers and other content writers, it's important to note that LLMs like ChatGPT can amplify social biases and produce false or misleading results. This is due to the questionable data included in some of the millions of websites that it is trained on. In conclusion, ChatGPT can be a useful tool, but its results should always be validated and verified before being relied upon.