AI solves for inefficiency

I spent more than a couple of years writing tickets in jira. In fact I think I may have written over 500 detailed tickets. All of which had taken into consideration the current state of the UI as per the design team, the available API endpoints and the needs as expressed by the Product Owner. Let me tell you, it is not a fun job. And once you start down the path of writing Gherkin into something that satisfies all your corner and edge cases, then you're going to be refining and polishing for too long. You'll be writing a ticket a day.

It was this neverending repetition that inspired me to cull the content of my tickets to the bare minimum and to make them as useful as possible. This ended up in a ticket format that looked a little like this:

This format had well structured information relating to some of the behaviours expected from a feature, which both developers and business stakeholders could understand. I could write these very quickly and anyone could comment on them without difficulty mostly because they were not prescriptive about the how nor did they give the 'business case' for the feature. It was the feature, plain and simple.

Even so, after reviewing documentation and talking with individuals I might be able to get through 10 to 15 of these a day. No problem, that's adequate. And it continued to be for a long time.

What always got me however was having to start each one from scratch and I attempted many times to automate this. No luck. Until...

Just a couple of days ago OpenAI made custom GPT's available, I immediately decided to automate a personal task I was in the middle of. Transcribing a large number of recipes ripped from magazines. It took ages, and I had to read each recipe carefully and format it the way I wanted it. Suddenly I was getting a recipe into Notion at a rate of 1 a minute instead of 2 a day. All I had to to was take a photo of the recipe and paste it into the GPT. Recipe transcriber was born. Recipe automation for the win.

And then it ocurred to me. I could do this with tickets. And so Feature Ticket Generator was born. Now I can work through the creation of a ticket, have information elicited from me and have a draft ticket ready in around 3 minutes (or thirty seconds if I want to go the quick route).

In the first half of the year when I had the ocassion to have lunch with people I would endavour to ask their opinion on chatGPT and the consensus was "We're all going to lose our jobs". I'm now of the opinion that we are not going to lose our jobs. What is going to happen is that our jobs are going to become vastly more efficient.

We're not losing our jobs to AI, we're making them more efficient because of it.

Because workflows will be shorter the higher level question for organisations is: Do you want to increase throughput and keep your 10 people? Or, do you want to remove 5 people and keep throughput the same.

We won't lose our jobs. But we will collectively be able to do more, and that has consequences good and bad.