Story time: One of our dear customers told their epic tale of manually and meticulously crafting a list of 1000 keywords for their Auto-moderation rule. Yes, you read that right, 1000 keywords... manually adding them bit by bit over several months. π
But guess what? Those days are as outdated as VHS tapes.
Now, AI tools like ChatGPT can do the heavy lifting for you. Generate your keyword lists quicker you can say "NapoleonCat"!
Let's see how. π€
Requirements
βοΈ A ChatGPT account. If you're not on board yet - just sign up here (The Free Plan will do, but GPT 4 is superior).
βοΈ A Pro Plan with NapoleonCat to unlock access to Auto-moderation. Not there yet? Here's how to upgrade
π ββοΈ Remember, never share sensitive data with ChatGPT or any other generative AI tool.
βοΈ Thinking about other AI tools like Bard? They could work, but from our experience, GPT 4 is the star player in keyword generation.
That said, if you're on the free plan with ChatGPT - it should do just fine.
Now, let's see how to generate lists of keywords for your Auto-moderation rules without breaking a sweat!
Generate keywords
Whether you're handling product inquiries, battling crypto spam, or responding to discount requests, ChatGPT can swiftly generate your keyword list in minutes, not months!
Here's how to generate a list of Auto-moderation keywords with the help of AI:
Log in to your ChatGPT account
Log in to NapoleonCat and head over to Auto-moderation
Keep the NapoleonCat tab open for seamless navigation
Start a new conversation in ChatGPT:
ποΈ Time to get interactive!
Copy the following prompt and paste it into the conversation window, but remember to switch the Example Values with your specific needs:
Prompt
Say, for instance, you want to handle inquiries about "How to use your features." Then, you'd gear your prompt this way:
Examine the task before proceeding.
I need a comprehensive list of keywords for NapoleonCat's Auto-moderation tool, which is designed to moderate social media content automatically.
This tool checks for specific phrases or keywords in comments, posts, and direct messages. To get it running, I have to input a list of keywords, triggering the auto-moderation workflow.
We can use an asterisk * before, after, or both before and after a word sequence, and it helps apply rules to words with that specific sequence. Here are some examples for clarity:
1. Inputting *use* will trigger the rule for any words containing "use," like 'reuse' or 'usage.'
2. Inputting feature* will activate the rule for words starting with "feature", such as features, featuring, etc.
3. Inputting *ful will apply the rule to words ending with "ful", like useful, resourceful, etc.
4. Inputting 'guide' will activate the rule for the standalone word 'guide', and when it appears within a phrase, like "Here is a guide."
5. Inputting 'step by step tutorial' will trigger the rule for this exact sequence of words.
With this in mind, I'd like to generate a list of [INSERT YOUR NUMBER HERE] keywords for the Auto-moderation rule.
My goal is to respond to any comments that ask about "How to use our platform's features." and similar. [REPLACE WITH YOUR EXAMPLE]
I'd like to have keywords both in isolation (without an asterisk) and with an asterisk before, after, and on both sides of the word sequence, and phrases like "user guide."
It's important that you distribute the keywords roughly evenly, so that every scenario is covered. Keywords or phrases without asterisks are just as important as keywords with an asterisk before, after, on both sides of a word sequence.
Do not include common and generic words that could cause "false positives", and that are likely to trigger the rule when we don't want it, like "to", "he", "she", "for", etc.
β
DO NOT use the signs " ", or ' ' in your list.
Please separate the keywords with a comma, following these guidelines, and provide an extensive list. Format the keywords in the code format (in the code snippet).
Each keyword should be in a new line, but with a comma.
Result
π€ Results are in! After using the example prompt, ChatGPT served us this:
π‘ If ChatGPT doesn't nail it on the first go, don't worry! You can always guide it with additional prompts. Here's how.
Got your keywords all lined up? Brilliant! Let's slot them into your Auto-moderation rule:
Paste keywords
Go ahead and copy the list of keywords:
Next, hop into your NapoleonCat account. As you're creating or tweaking an Auto-moderation rule, find the "Trigger" section and paste the list there:
And voila! You're good to go.
Complete your Auto-moderation rule, and you're set to handle questions about "How to use a new feature in your app." π
Tweak keywords
Not totally happy with your list of keywords?
πββοΈ Generative AI is just a tool, and it's not flawless. You may need to roll up your sleeves for a little manual work.
Here's what you can do:
Review your keyword list and make any needed tweaks manually
Guide ChatGPT with some more prompts if something doesn't feel right. Not enough keywords with asterisks on both sides? Try this prompt:
You haven't used enough examples with asterisks on both sides. Please provide me only with a list of relevant keywords with asterisks on both sides.
βοΈ So, the AI-generated keywords are a great starting point for your Auto-moderation rules.
But, they might need a little tweak to suit your unique needs. Your community may use language in a slightly different way.
And what about "false positives"? Well, they're instances where keywords trigger Auto-moderation, but the content was actually okay. If a keyword is causing too many of these mix-ups, feel free to adjust or remove it. Your moderation, your rules!
That wraps it up! Who knew workflows could get a futuristic upgrade, right?
π¨οΈ Need a hand? Talk with us on chat in the bottom, right-hand corner of your screen.