07/09/2025
If it feels like every tool out there is suddenly calling itself “AI,” you’re not imagining things.
But here’s the catch, not all AI is built the same, and not every platform is actually helping your operation.
We just published a quick breakdown on what real industrial AI looks like, and how to separate the helpful stuff from the hype.
Is AI Good for my Business?
Written By: Payton McDaniel
In recent years, AI software has begun to take over aspects of many different industries. Every major company, such as Google, Amazon, and Facebook, has added an AI feature accessible to consumers. These AIs all claim to be something different, when in reality, they are large language models like ChatGPT, marketed as if they are a newer and better version of AI. Large Language Models are trained on human language and text using large data sets from libraries and the internet. Through individuals and corporations using AI, along with the information and writing styles it gathers from all corners of the internet, it learns to process written inputs more effectively. Language-based models(LLMs) can be useful in marketing and advertising departments, but this type of AI doesn't translate everywhere. Platforms like ChatGPT, Bard, and Sora are helpful, but they are generalists, designed to be competent at all tasks without specializing in any one area. Industries such as healthcare, law, and hardware automation, which require unique specialization, also require specialized AI. AI products are trained on tasks; for LLMs, it is language, but for others, such as Universal's Neocortex, it is robot path planning.
Universal Logic's Neocortex is an AI software that's trained differently. Instead of being trained on information on the internet, it is trained on understanding variable 3D environments and planning robot path movements. What makes this software special is that instead of having a roboticist hardcode XYZ positions for the robot to follow, Neocortex uses 3D imaging of the workstation to identify any potential obstacles or changes in the pick area. Then, in real-time, the AI software makes decisions on how to pick the object that both the most optimal path is considered while also taking into account other data, such as the orientation of the object or any potential hazards that could arise. After running the AI over 500 million cycles in the field, there's almost no pick software that hasn't been trained for. The software can be easily paired with any style of robots or sensors, making its use applicable for almost any automation task.
Leveraging parallel processing, we can make decisions faster than traditional systems and simultaneously collect and analyze multiple sets of information to expedite the decision-making process. For example, a box sorting project with four different colors: black, purple, blue, and orange. An operator can instinctively identify and sort objects based on colors, but automating this identification process presents challenges for traditional automation. This type of project needs quick thinking to differentiate colors. With imaging and sensors used in conjunction with AI, the software can make timely decisions on which path the object presented needs to follow. This software can successfully complete the sorting task faster, leveraging these capabilities.
AI is a tool that is privy to how you use it. For some of your more mundane administrative tasks, ChatGPT could be beneficial. Trying to automate your palletizing? Perhaps a more advanced, industry-specific AI is a better fit. Remember, just because something is branded as AI, it doesn't necessarily mean it's different from everything else. Do your research when it comes to the software and the platform, and when you come across a Neocortex, think of all of the possibilities.
