Enterprise Applications slowly adopting AI development

Many companies have started changing their development strategies. Initially when AI was in the market, companies were scared to let their employees use AI. My company had severe restriction on AI usage during the early days. I used to use Chat GPT on my mobile to solve programming problems. But later the companies started understanding the importance of AI and slowly brought the adoption. The AI companies also started providing private enterprise solutions where the data will not be sent out anywhere. That was the biggest selling point where it worked very well for the companies to adopt it.

Source - Gemini

Due to lack of awareness, only a limited amount of people started using AI in the beginning. I personally used it only for some of my questions or for brainstorming a solution. Later when AI improved and started providing code snippets as solutions, it started becoming interesting. I used a lot of AI to get code snippets for my coding needs. Then later Agents were introduced. But even before that AI would review the code and proide just suggestions and we have to implement those suggestions manually. The next best part was when AI was able to make code changes automatically. This drastically changed the way coding happens.

Today we have reached a transition where in many teams the developers are asked to use AI to build and it has even become mandatory. Some people like it because it saves a lot of time. Now for the past few months I have been heavily using AI for most of my needs and I'm not even sure if I will go back to manual coding at all. Even for a small task, I get it done only with the help of AI. My work is becoming heavy only with the reviewing part where I have to review the code written by the AI.

The biggest challenge for AI assistance in Enterprise application is that those applications will have a huge codebase and if a developer relies on AI too much, it can end up disastrous. When AI is not given the right context, it gets confused and at the end it gives incorrect solutions. Anyone using AI for bigger code base should be very careful because AI cannot be precise. We have to either split up the code and convert it to multiple services and then make use of AI or be very careful with what we are asking the AI to do. But this is just the current status but in another few months AI world will have rapid changes.

If like what I'm doing on Hive, you can vote me as a witness with the links below.