The recent AWS outage gives the context to share a story from Q1 2025. I was working on defining cloud application development guidelines for the org. I resisted the activity initially for few reasons:
- All the guidelines you ever want is already available on internet. Why spend more time on it?
- More difficult than writing guidelines is having people follow it.
- It’s hard to define success criteria and measure impact for such initiatives.
The time was short (as usual) so I decided to focus on essentials and what mattered to our org.
-
I used Perplexity for my research because credibility and citations of original sources was non-negotiable.
-
Instead of looking for generic guidance, I directed my search to specific problems we had faced and that we wanted to avoid. I structured the guidelines in a goal oriented format. For ex. How to connect from my app to managed services on AWS without exposing the traffic over public internet?
-
For each section, I provided concrete check lists relevant to our organisation. (The ideal approach would have been to automate these checks for compliance but I thought of getting there in stages.)
-
I gave rationale as well as reference to guidelines from reputed sources to drive the point home that these practices were adopted industry wide and not following them would make us an anomaly.
-
I referred to best practices from multiple leading vendors to look for universal principles.
In a couple of weeks, the guidelines were in place.
Every single line written was scrutized for correctness, valid reference and relevance.
Did AI save lot of time as compared to Google search?
Not much.
But it did make the effort less boring and more interactive and fun.