As we get into the last quarter of 2024, it has gotten to a point where there is so much noise about AI that it’s difficult to discern what is actually worth paying attention to. Every major enterprise software vendor has “AI” embedded into their product. While GPT has been an incredible leap forward, its actual ROI hasn’t been widely realized. There are several reasons for this, which I will get into later. We, however, are already seeing some fallout from executives who made reactionary spending choices in late 2022 and 2023, where they wanted to bring AI into their organization but didn’t have clear ROI, were overpromised functionality, and the expectations didn’t meet reality. This is very common with hype cycles in tech; there will always be early adopters who want to be the “first to do it” and add it to their resume. This can be dangerous thinking based on the high costs of AI to get it to work properly.
So where are we?
It’s a good question. I continue to contend, as do many others, that we are in the equivalent of the early-stage proliferation of the Internet as a GUI and interactive application. Think Mosaic browser and Yahoo Search engine, circa 1993-1995. Many of those reading this most likely will not have been born yet or even know what I am referring to. Let’s just say this was a period when e-commerce was still a few years away.
To compare the eras is doing the capability and promise of AI a bit of a disservice. AI has made the creation of AI more efficient, so in the case of the Internet era, it would take years for improvements to be made, i.e., browser changes, website functionality, HTML version changes, etc., which has helped advance the Internet from simple block text to today’s applications which are highly interactive and at greater scale. What we are seeing are major advances in the usefulness of AI and at breakneck speeds. GPT is everyone’s favorite to look at, but advancements in Multimodal AI, which can process various types of data simultaneously (text, images, unstructured data, etc.), open up a whole new way in which we as humans can use this tool to assist us with our day to day lives in our personal and business lives.
This is all very exciting, but I would encourage all of us to take a half step back and assess where we came from and where we want to go. At the end of the day AI is a data problem. The more QUALITY data you feed it the smarter it can become. That statement alone eliminates most workflows as the data can’t be had or there isn’t enough to train a model. As a company, we are in the process of building very sophisticated AI models that solve a problem, while it may be large and high in ROI if it is solved. The approach is that it can focus on enough historical data that produces an outcome that can be analyzed. More to come there! In closing, the future is going to be great with more AI in our lives. I advocate you to read through the lines and choose wisely as you proceed on your journey.
Written by: Peter Camp, CTO & Founder
