AI: Your Best Intern Ever

Everywhere you look, every event we attend, every vendor update seems to include something about Artificial Intelligence (AI). Some of it sounds fluffy, like they have plans, but nothing is available yet. Others promise the world if you use their iteration or system! Chris suggested interviewing our teams and sharing some things they do with AI. We hope to inspire you and help get the creative juices flowing!

I met with Mariano on our Developer team. He’s been with us since early 2018, so he’s been part of our AI journey for a while now. AI isn’t new to this team. The Devs have been using it mainly for coding for years. These tools are very helpful when writing and deploying code, as well as finding errors and checking results (testing). In the last 18 months, they’ve found some ‘favorite’ tools and really honed in on how and where to use them for the most impact. Using the right tool makes all the difference. The really interesting part is how fast the use of AI has accelerated since Copilot became so readily available.

Internally, our Dev team has provided some pretty amazing solutions to help Syscon and our clients stay streamlined, standardized, and bullet proof. With our Daily Field Reports add-on for our FIT System, AI helped with the mock testing and the very time-consuming copy functions. With a little fine-tuning by the Devs to address small issues, the project was much more efficient (and fun). We use a ticketing system and our FIT System for time entries. The team was able to embed AI into our FIT System to generate alternative wording, find spelling errors, and similar. They even have a link back to the ticket, right from FIT!

Our Business Applications (Biz Apps) team often uses AI tools to work through and build the framework based on the project or concept for a project. This outline process has made them much more efficient as they dig into the project. When you ask Copilot to talk with you as if it’s an expert, the feedback and ideas are excellent. Sometimes we have more than one tool or path available for the project. Using AI allows this team to efficiently think through the pro’s and con’s for both options. When it comes to syntax, Copilot can handle as much as 90% of the draft, allowing our team to take it from there, saving a lot of time (and drudgery). I’ve heard several of them turning to AI when they feel stuck with the code or project, and many times, the suggestions have helped spark how to move forward.

Everyone commented on using Copilot for internal meetings! These provide great summaries and it’s easy to return to the meeting to re-listen to a section, or just ask Copilot specific questions about what was said. Copilot also follows our SharePoint security rules, so our data is safe and secure.

Mariano feels the whole Dev team is more efficient using these tools. They’ve taken the time to focus on repetitive tasks (their least favorite), figuring out which tool is best, then ‘teaching’ it what to do or look for. By treating the AI tool as an intern or trainee, the Devs can delegate the boring tasks to their AI models, then, just like with a human intern, provide supervision and check the results. Since the AI tool is learning over time, its results get better and better. I’ll bet you can make a short list of repetitive tasks; which ones might be worth trying this on?

Many of us are familiar with the amount of time it takes to keep procedures up-to-date, vendor lists, and similar data that’s been around a while. In the Development world, it’s old libraries. These are time-consuming to update. With Copilot, the Devs were able to find the necessary changes from the old to the new libraries. This kind of work would have taken a week just a few years ago. The last round of updates was about six hours! With this kind of time savings, they’re able to do updates more frequently; a win for them and for our clients!

It does feel like AI is moving very fast, faster than this team can stay ahead of. These tools seem like magic! Initially the Devs were skeptical. They didn’t think the tools could write code successfully. After digging in and treating it like an intern or trainee, assigning the right tasks and providing the right direction, it really did feel like magic. When learning a magic trick, the first couple of times you probably don’t get it right. But with repetition, you can be great; same here! – CMW