My First Steps Toward a Daily AI Process -
James Land - 05 Sep 2025
My First Steps Toward a Daily AI Process
As I am sure many of people others have already figured out, this year AI has moved from being a novelty to really becoming a game changer.
I have been using AI for a few years now with GitHub CoPilot and it massively accelerated my development process. But recently I tried to up my game with other tools such as Windsurf and Cursor.
The output that I am able to get out of these LLMs have been frankly quite impressive. Espically when dealing with librariesy I am less familiar with, I have found that it has helped me much more quickly understand how to use them and come up with working solutions.
The Goal
Because AI has been so helpful in my coding workflow, I want to find a way to help make it a part of my daily workflow. And use it to help me with daily task.
Specifically with this article I am going to focus on more effectively asking AI questions in a fast and more organized way.
My Current Process
Right now when I want to use Gemini I generally open up a new tab and paste in my question. It is clear to me that even though my questions include little context that Gemini processes them to create a better prompt for itself internally and still returns a good response.
While this is fine for most of my questions, I would like to be able to give stronger prompts more consistently to get better results.
The Proposed Process
There is the idea of “Gemini Gems” which I have played around with and they work great. But they cant take time to construct and you have to remember to use them.
I found a suggestion on line of using a “Text Expander” to save a set of custom prompt templates, that I can quickly edit and use to give better results.
I have never used a Text Expander before but it seems like a solid idea so I found Espanso which works with Fedora (a little tricky on Wayland but not too bad).
Using Espanso
Now that I installed Espanso I need to create a set of custom prompt templates. So I am starting of basic.
Being a Red Hat employee I use have used AI Chat Engines to ask questions about Red Hat Products, especially around Red Hat OpenShift. One issue I found is that I often get information that applies to a different Version of the product than the one I am currently using. So lets create a prompt template that will give me the information I need for the version of the product I am currently using.
First I am going to open up the the ~/.config/espanso/match/base.yml file (using the command espanso config) and add my prompt:
- trigger: ":ai-rhocp"
replace: |
# Red Hat Openshift Question
* **Goal:** [Help me understand these logs]
* **Version:** e.g., OCP 4.14 on AWS IPI.
* **Problem:** What's happening vs. what you expect? Include any error messages.
* **Base Domain:** [apps.rosa.jland.guid.ab.openshiftapps.com]
```yaml
# Paste relevant YAML here
```
Now when I want to trigger my prompt while inside of Gemini I can just type :ai-rhocp and it will replace it with my prompt.
Next Steps
Now that I have created a basic prompt I want to expand to create other prompts for different tools. For that I am going to save my config file in a separate git repository that can be found here, and will hopefully grow over time.
I even AI’ed up a little workflow so I can create my prompts in separate files and then combine them into a single config file automagically!
Conclusion
This is my first foray into using AI to help me with my daily workflow. I am sure there are many ways to improve this process and still a lot of stuff to learn, but I hope to keep up with this and try to document my journey here incase anyone else is interested in doing the same.