Customer Success is always under pressure to optimize (and reduce the expense of) everything. Many people see ChatGPT as a way to streamline Customer Success processes like preparing for customer meetings or literally creating those processes upon which their organizations run.
One approach some Customer Success pros are taking is to use ChatGPT to generate Customer Success playbooks. ChatGPT, if you’re not aware, is a language model trained by OpenAI that can generate human-like text and help create high-quality content.
Let’s dive deeper into how businesses are using ChatGPT to create Customer Success playbooks, examine the potential drawbacks of doing this, and provide tips on how to use ChatGPT effectively for this purpose.
The Benefits of Using ChatGPT for Customer Success Playbooks
One of the primary advantages of using ChatGPT to generate Customer Success playbooks is the time and effort it saves, and the expense of buying (and continuing to pay the high cost of trying to implement) generic “blueprints” from consultants.
With ChatGPT, you can essentially avoid starting from scratch and generate what appear to be high-quality playbooks, quickly.
However, there are some limitations to using ChatGPT that can seriously impact the quality of this output.
The Drawbacks of Using ChatGPT for Customer Success Playbooks
First, the underlying model of ChatGPT was trained up to mid-2021, and therefore does not include the latest industry updates and best practices.
Customer Success is continuously evolving, but has evolved rapidly and drastically over the last couple of years. This is where the Bing Chat + ChatGPT strategy helps, because Bing will pull in live search results (where ChatGPT doesn’t have access to live web data).
It’s also critical to remember that the vast majority of Customer Success content is written by people that do not understand Customer Success.
This isn’t a dig on so-called “experts” in this field, but literally, staff writers and content teams do “research” and generate content on this subject – and many others – that rank well (I see you Hubspot), but don’t add anything new, or at worse, perpetuate outdated and disproven (or simply irrelevant) ideas.
So much Customer Success content is simply derivative of a few original (and long-outdated) sources, or derivatives of derivatives, and so the same nonsense just keeps getting proliferated.
This hasn’t changed, BTW, and with ChatGPT and other generative models, this is only going to get worse. I am fully aware.
Anyway, that derivative, outdated, but authoritative content is made more authoritative as it’s the basis for content on lower-authority sites, and thus the vicious cycle continues.
So when ChatGPT was “reading” the web, the connections it made in it’s model about what Customer Success is, how it works, etc. are skewed by all of that noise.
Finally, and critically, the ChatGPT-generated content will lack context about your company, product, customers, Customer Success org, Coverage Models, and the unique relationship between all of those things.
Examples of Using ChatGPT for Customer Success Playbooks
To better understand how ChatGPT is being used to create Customer Success playbooks in practice, let’s examine a specific example.
A SaaS company that wants to improve its onboarding process can use ChatGPT to generate an onboarding playbook. The great thing about ChatGPT is you could literally just ask it to do exactly that – generate an onboarding playbook – and it will. Like magic.
But is it useful? Is it specific to your situation? Is it based on up-to-date information?
In that case of this very simple prompt, no. Not useful at all.
See, you can either look at ChatGPT as a magical machine that conjures content as if from a separate realm or you can understand how to use it as a serious business tool and get it to actually create useful output for you.
The latter, by the way, takes an extra few minutes – max – if you know what you’re doing. And those who buy our Prompt Engineering for Customer Success course know what they’re doing.
Training ChatGPT for Effective Customer Success Playbooks
To ensure that ChatGPT-generated playbooks are accurate and effective, businesses need to train the model with specific information related to their unique situation.
To get something usable in reality, you need to input relevant information and data related to each stage of the onboarding process, differences from one customer segment to another, the Coverage Model for each segment, etc. and then ChatGPT can generate a playbook closer to what you need.
Then you can refine from there.
This training process includes feeding the model with up-to-date information and current best practices, as well as using specific prompts that provide context on the customer, market, industry, and other relevant factors.
That training should be followed by actually taking advantage of the chat aspect of ChatGPT to use refinement prompts to shape the content until you get what you’re looking for.
We have an entire section in Prompt Engineering for Customer Success on how to effectively train ChatGPT.
ChatGPT is a Powerful Tool (If you Use it Correctly)
Using ChatGPT to create Customer Success playbooks is an innovative approach that can save time and effort while providing fresh insights into existing strategies.
By training ChatGPT with specific information and using relevant prompts, Customer Success pros can create playbooks that are actually useful.
Unlike tools of the past where “Garbage in, Garbage out” was the mantra, ChatGPT obfuscates the intake of garbage, and it’s magical generative powers make the output that might be garbage, seem amazing.
The hype is real, but the usefulness is only there if you understand how to use it.
About Lincoln Murphy