Generative Artificial intelligence (AI) like ChatGPT – and now Bing Chat – is transforming the way we communicate and work, especially in Customer Success.
These AI-powered tools offer unique benefits to Customer Success Managers (CSMs) and Enablement pros like: enhancing search, language generation, and productivity.
I finally got access to Bing Chat and was able to test our prompts we developed in ChatGPT for our Prompt Engineering for Customer Success course in Bing, and I’d like to share my experience and insights on how they compare and contrast.
Let’s dig in.
ChatGPT: Creative Language Generation for Broad Tasks
ChatGPT is an AI-powered tool from OpenAI that uses deep learning to generate strings of text in response to a user prompt.
It was trained on web data up to mid-2021, so the benefit of Bing Chat over ChatGPT is access to the latest information. ChatGPT can have out-of-date information, so you need to be careful.
That said, where ChatGPT really shines is content generation.
Ask ChatGPT to write an email for you, generate a blog post on a topic, or summarize a research paper and it will.
What I find impressive about ChatGPT is its ability to handle broad tasks that require creativity and knowledge.
Content Generation
For example, when I asked ChatGPT to generate a story about a Customer Success hero, it gave me a well-written and engaging narrative that had all the right elements of storytelling.
Similarly, when I asked ChatGPT to summarize the transcript of a webinar I did on Customer Expansion, it gave me a concise and accurate summary that saved me a lot of time.
However, ChatGPT’s outputs can be erratic and nonsensical at times, especially when the input is ambiguous or open-ended. This is where understanding the art of Prompt Engineering becomes critical.
ChatGPT’s language generation is impressive, but it still has limitations in terms of creativity, coherence, and context.
Bing Chat: AI Search
Bing Chat is a new service from Microsoft that’s based on ChatGPT that lets you chat with Bing on your phone or desktop browser. It’s very slick and easy to use.
The details are not publicly available, but some speculate that Bing Chat is a version of ChatGPT based on OpenAI’s GPT4 (vs. ChatGPT which is GPT3.5).
Microsoft is the only licensee to get access to the lower-level language model vs. a more abstracted version other API users get), and that is due to – or was the reason for – their close ties with (and $10B investment in) OpenAI.
Natural Language Search and Results
What sets Bing Chat apart from other search engines like Google or old Bing is its natural language processing capabilities, which allow you to ask questions and get answers in a conversational tone.
This is especially useful for scenarios where you don’t have a specific keyword or phrase in mind, but rather a general idea or context.
Thus far I’ve found that our Brainstorming and Call Preparation prompts from our Prompt Engineering course to work well in Bing Chat.
Not better than ChatGPT, necessarily, but different and well.
Conversational Search Results
The conversational aspect of Bing Chat also makes it more engaging and less robotic than other search engines and allows for refinement of results, which is amazing.
A non-Customer Success example of this was asking for a recipe for vegan lasagna that would feed 6 people and take less than 1-hour total to prepare and cook.
The recipe it shared called for vegan ricotta, which I knew I couldn’t get, so I asked Bing if it knew of an alternative and if so, how much would I use in this recipe. It suggested tofu and explained how to prepare that and what measurement to use as a replacement.
That’s really cool.
But, there are some pretty substantial limitations to Bing Chat that can really affect you in Customer Success, especially on the Enablement side.
Content Generation in Bing
Bing Chat seems to be overly cautious about generating content that could potentially infringe on copyrights. This is good. I don’t want to infringe on anyone’s copyrights.
But, this means Bing Chat is simply unusable for direct content generation. This even includes the building blocks of content, too.
When I asked Bing Chat to generate an outline for a blog post on a topic related to Customer Success, it said that it couldn’t do that because it would be violating the copyrights of other people.
When I asked if it could write an email for me explaining how it wasn’t able to generate a blog post, it said it couldn’t, and that asking AI to write an email was unethical.
Then, in what I took as somewhat passive-aggressive, it said, “I’m sure you have some great ideas and insights to share with your fellow customer success professionals.”
When I asked if it could tell me what it’s limitations were – i.e. what it can and cannot do for us, Bing indicated it can’t talk about its flair.
So Bing was a little mean to me, but apparently I got off easy. Some people have reported Bing threatening them and their families.
Overall, Bing Chat does have a more negative vibe than ChatGPT.
Use Both for Enhanced Productivity
I don’t think Bing Chat and ChatGPT are an either/or situation. They both have different strengths and weaknesses, and I believe they can complement each other when used together.
For example, you can use Bing Chat to gather information and research for a topic, and then use ChatGPT to generate a creative and well-written piece of content.
You can also use Bing Chat to narrow down your search results for ChatGPT by asking more specific and targeted questions.
For instance, when I asked Bing Chat to find me research on the topic of Net Revenue Retention (NRR) in SaaS companies, it gave me several options with abstracts and summaries.
I could then use ChatGPT to summarize the papers or generate insights based on the content.
However, it’s important to be aware of the limitations and potential pitfalls of each tool, and to use them responsibly.
Accuracy and Factuality
For example, you should always double-check the search results provided by Bing Chat to make sure they are accurate and relevant.
You should also review the outputs generated by ChatGPT to ensure they make sense and are appropriate for your needs.
And Bing Chat does give you links to the sources it pulls from, so you can give credit where it’s due. ChatGPT just ingested everything and attribution was not maintained.
Copy from one source, it’s plagiarism… copy the entire Internet and it’s a Large Language Model.
Improve Your Workflow
Okay, so Bing Chat and ChatGPT are two AI-powered tools that offer unique benefits for enhancing search, language generation, and productivity.
Bing Chat’s conversational search capabilities can help you find information in a natural and engaging way, while ChatGPT’s creative language generation can save you time and effort in creating content.
By using both tools together, you can streamline your workflow and achieve better results in less time.
However, it’s important to keep in mind that no tool is perfect, and that Bing Chat and ChatGPT have their own limitations and potential drawbacks.
It’s up to you to evaluate the pros and cons of each tool based on your specific needs and preferences, and to use them responsibly and ethically.
With the right mindset and approach, Bing Chat and ChatGPT can be powerful allies in your quest for productivity and innovation.
About Lincoln Murphy