Author's): Felix's carton
Originally published in Towards Artificial Intelligence.
AI ChatBots have already been integrated into many of our lives, and if you haven't done so yet, perhaps this blog post will help you start using them to free up time from tasks you don't enjoy.
In many situations, I used large language models every day for the same tasks, such as correcting spelling in English, learning a third language, summarizing long texts that I don't have time to read, or brainstorming project ideas.
However, always starting from scratch with an empty tooltip is time-consuming and inefficient.
Sometimes writing a decent prompt takes even more time than completing the task itself.
The solution to this problem is automatic suggestion.
Instead of writing the same prompts over and over again with different results, you can create innovative prompts with reusable instructions that you can improve over time. This improves the quality of the results and saves time and effort.
In this blog post, I share my thoughts on building automation prompts, how to continually improve them, and what to keep in mind when creating such prompts.
What does automation suggest?
Automation prompts give AI models a clear set of instructions to reuse.
These instructions can be run manually or automatically based on external triggers such as a message or at a specific time.
This post will focus on manually executing automation prompts because we need to first understand the process and execute it manually before automating it.
The idea behind auto-hint is that you write a smart prompt once that leads to the result you want, and then you reuse it and continually improve it.
Even though it may take many hours to create such a tooltip, it is worth the time investment in the long run.
At any time in the future, when you encounter the same problem again, you can copy and paste the prompt and avoid starting with a blank prompt input field for repetitive tasks.
The inconvenient truth about automation prompts
The secret to automatic tooltip generation that no one wants to hear is that you must first complete the entire process manually before defining an effective tooltip.
Once you've done the process a few times, you gradually refine it.
All the knowledge you gain from performing the task repeatedly should then be integrated into the automation prompt.
The key to an effective system message is to effectively translate experience into clear instructions.
However, it is easy to mistakenly provide too much detail when trying to verbalize everything in your head.
Imagine your LLM as an intern at a large company. If you want to successfully guide an intern through the enterprise, you need to provide detailed instructions.
However, too much information will confuse the trainee, preventing them from acting effectively.
There will always be a trade-off between providing too much and too little instruction.
The same principle applies when working with AI chatbots.
Automation prompt example
Let's take a look at one of my simplest automation tips that I use every day and maybe you will soon too.
There is always a hint phase at the beginning.
As someone with a keen interest in artificial intelligence, I actively try to stay up to date with the latest news.
However, if you've tried it yourself, you've probably noticed that it's a daunting task due to the daily deluge of information, product launches, and breakthroughs.
Faced with this challenge, I had a choice: give up or turn the task into an automation prompt.
Since I had already done this process many times, it was easy for me to condense the most important insights into a concise form.
You can see this prompt below.
What are the latest developments in artificial intelligence?
Divide the output into the following sections:# Research and breakthroughs
New articles, model publications or noteworthy academic discoveries.# Industry and product news
Announcements from major technology companies, startups or product launches.# Politics and ethics
Government regulations, ethical debates or updates on global AI governance.# Social impact
Stories about how artificial intelligence is impacting workplaces, education, media and everyday life.# Notable quotes or comment
Key insights from thought leaders, CEOs and researchers.”Write a paragraph of at least three sentences for each message you find. It's okay if you don't find anything for a given category.
Now I use this prompt every day to gain momentum in the world of AI.
You can see part of the result here.
Of course, it's not as detailed as manually searching for the latest news on the Internet.
However, this generation took me less than a minute and I can even listen to it.
This way I turned it into a more sustainable habit instead of giving up and doing nothing.
What's more, if you want to listen to the recordings in a more entertaining way, you can turn these messages into your own podcast episode.
This can be achieved with the next prompt.
A drawing robot automating the output of tasks with marked boxes on a conveyor belt titled “Automation prompts”
Some of the results of this prompt are shown in the next image.
This is the first example of an AI prompt chain and workflow.
In such a workflow, you take the output from one model and use it as input for the next one.
However, it does not have to be the same model.
For example, you can even extend this prompt automation chain by copying your podcast script into a text-to-speech model like ElevenLabs, which has shown amazing results with the latest V3 update.
Let's assume that a manual prompt automation workflow that consists of different prompts improves over time as the quality of those prompts improves.
In this case, you can start automating the entire process using tools such as Make, n8n, or Zapier.
However, don't rush this process. Many rapid automation projects fail because people try to automate something they have never done before manually.
Tips and advice
Tip 1: Don't have too high expectations for the first version of a prompt
Don't expect a perfect prompt template from the start.
Fast automation starts with the simplest template possible and gradually improves, adjusting step by step to better fit your needs.
Japanese rule Kaizen can help you when creating a set of automation prompts.
It stands for “continuous improvement” and focuses on making small, incremental changes over time.
I'm always thinking Kaizen while experimenting with my prompts.
Tip 2: Write down your prompts in one place
Be sure to write down your prompts somewhere. I personally use Notion to store all my prompts, but any other note-taking app works too – such as OneNote, GoodNotes, Notability, etc.
Tip 3: Leverage your copy history
Did you know that Windows' copy and paste feature does more than just insert the latest copy?
Urgent Windows key + V opens the Windows clipboard history.
If you've never used this option before, Windows will ask you to turn it on.
This feature saves the recently copied items and allows you to select the one you need. You can even pin your most important copies. I use this to pin my latest prompts so I can access them with one shortcut.
Unfortunately, as far as I know, macOS doesn't have such a feature built-in.
You'll need third-party apps like Paste, Raycast Clipboard History Extension, or Maccy.
Last words
Think of your tips as tools: the sharper you sharpen them, the better they will serve you. Over time, well-crafted prompts can save you hundreds of hours.
But where should I start? Start small. Choose one task you repeat often and write your first automation prompt for it today.
Don't worry about it being perfect. The magic is in iteration.
Just get started and improve as you go.
We are all learning this, one prompt at a time.
If this post helped you, please share it with someone who writes the same prompt from scratch every day.
Help them turn repetition into automation.
Follow and connect
felix-pappe.medium.com/subscribe 🔔
www.linkedin.com/in/felix-pappe 🔗
https://felixpappe.de🌐
Published via Towards AI
