Overview
When consulting with clients, one of the most requested tasks for us has quickly become the integration of Artificial Intelligence features. Potential clients either know exactly what they want and why, or they have some general idea that AI might improve their software making it or their company more efficient in some way without any specifics.
The Three Questions You Need to Ask of Every Component Of Your New and Existing Business Software
We’re of the belief that finding the best outcomes with AI doesn’t start with asking where you can add AI, but rather with these three questions:
- Do we have problems previously impractical to solve that are opened up with AI?
- Does AI improve upon our existing solutions to existing problems?
- When encountering a new problem, is AI the better solution?
The answer to any of these questions is often no, which might come as a shock to a business world that’s moving so much money around to invest in AI that it makes the entire United States defense budget look modest in comparison.
Despite the overhype, as long as you don’t get caught up in the mania and shove your business into being ‘AI Forward’ at any cost, the potential benefits are real. Most businesses will have uses for it somewhere. Not being left behind doesn’t mean shoving AI anywhere it will fit, it means not delaying a careful audit of your business practices and the software you are built on top of and asking the AI questions to each process and component.
We’re going to look at two case studies of how we used AI for internal projects. One to add limited AI features to an existing software product, and the other to build a software product from scratch.
A Light Touch: Adding AI Features To Simplified Hire
CodeClouds recently added AI to one of the software products we offer under the BeSimplified brand, called Simplified Hire, an applicant tracking system. We know too many recruitment systems are using AI too heavily, letting it be the only way to get on a recruiter’s shortlist. Instead, we architected our AI Fit Score feature to move up the candidates who are more likely to be the ones you need to contact first lest you lose out on them to another company. The interface is designed in such a way that AI never filters out candidates, it instead promotes who to review first and assumes that eventually, a recruiter will work at reviewing all of them until the role is filled.
We also use AI to assist with document management on the platform. Too many candidates have had to format their resumes and CVs to fit into narrow fields in larger companies’ ATS systems so they can be acted on by automation. This is one of the problems where AI has an objective direct benefit over the old way, allowing it to parse documents of various formats and structures to fill in the summary information that goes in the recruiter’s dashboard.
There are still many features in Simplified Hire that don’t currently use AI. For each of these, if we see an objective benefit, we will add them, as well as consider whether AI is the right choice for each new feature we add.
Heavier AI: Saving Development Time on Simplified Spaces
In the case of Simplified Spaces, another software product we offer, we saw fit to use AI in the development process, but didn’t feel the product itself needed AI features as of yet. We needed to fast-track the development of a new software product for managing coworking centers, and we didn’t have enough engineers to start immediately to take on a project of this size on short notice. One of the critical time issues we were facing was that another partner business of ours needed custom software to run internally as soon as possible. We decided to develop this product using both AI and our developers, targeting having the majority of the code be AI generated. For the final product, around 75% of the code was made with AI agents.
Using this strategy, we built a product that would normally take us more than a year with the team that normally works on software development for BeSimplified in a couple of months. It wasn’t as simple as telling AI what we wanted and generating a software product. We still had to design and architect the product we wanted, down to the details of how each feature was meant to work. After new code sections were created, our engineers had to review them, test them, and squash any problems. Our engineers both knew what the end result was supposed to look like, and what sections of code were better suited for AI generation, and what sections of the code would likely end up taking more time fixing what AI made instead of just making it by hand.
The difference between some person AI generating a software project versus a team of experienced engineers using AI carefully is night and day. In the right hands and on the right project, a lot of time can be saved by going almost ‘all-in’ for AI-driven projects. Even in this case, where most of the code was AI, it wasn’t a matter of days, simply letting AI agents work in the background eating tokens until a product popped out- it still was weeks of careful, intentional work. Weeks of work may sound like a long time for AI-driven development, but the end result is something that was actually solid and still got us over the finish line in weeks instead of months.
We’re Not Just Another AI Company, We Have Real Development History
We have more than 16 years of experience as a software company, and we’ve been working with AI since the breakthrough modern LLMs brought everyone. We can answer the questions of when, where, and how to use AI in your next software project. We’ll both encourage AI when it’s the right tool to solve a problem, and let you know when it may not be. It might just be a small feature made better by or not possible without AI, or it could be integral to the whole software development process, much like the internal projects we talked about in this blog.
If you have a new software project and you need to know more about the practical feasibility and timeline of development with heavy AI generation, or if you have an existing project and are interested in new AI features, reach out to us. Our interest isn’t selling you on AI when it’s not the right choice, we’ve been in traditional development for a long time after all. Our interest is in continuing our track record of successful software projects using the best tools for the job and advising you on the right features for your project.
How can we help?
Tell us about your inquiry and we’ll get back to you as soon as we can.
Kolkata (HQ)
Fort Wayne
Wellington
Sydney