There Is No Finish Line with AI

By: Andre Cliff

One thing I’ve learned from working very closely with AI is that there really isn’t a place of arrival. You don’t reach a point where the right model is in place, the right information has been added, and the work is done. There’s always something to test, adjust, clean up, or rethink.

That surprised me.

I used to think about AI more in terms of getting enough good information into a system and giving it the right instructions. These things matter, of course, but actually using AI in a business has taught me that there are many moving parts involved in supplying the information and getting a useful result from it.

Sometimes the information AI needs is already there and it still doesn’t get where you need it to go. How that information is organized matters. So does how the AI is directed toward it and the workflow built around the process. You can spend a lot of time trying to come up with the perfect prompt only to find that the real solution requires changing how you’re working with AI in the first place.

There’s some trial and error involved.

In my role as an AI Content Curator and Administrator, I’ve experienced this while using AI to help with work that involves curating information. What appeared to be a straightforward process took consistent tweaking before I found a workflow that could produce desired results. It wasn’t necessarily the textbook way I would’ve mapped it out beforehand, either. It was the way that worked.

I think businesses have to account for that. General-purpose AI already knows a lot, and the models are only getting more capable. But AI knowing your industry isn’t the same as knowing your business. It doesn’t automatically understand the experience your organization has accumulated, how your people work, what matters to a particular client, or which piece of internal knowledge should carry more weight in a specific situation.

That’s where curation comes in, and it’s also where people remain important.

There’s a familiar saying that knowledge is knowing a tomato is a fruit, while wisdom is knowing not to put it in a fruit salad. I think that describes the human role in AI pretty well. AI can process more information than any person reasonably could, and it can do it incredibly fast. A person still has to look at what comes back and understand whether it makes sense for a given client situation.

That judgment makes a major difference.

Clients are different, and what works for one may not work for another. AI can sometimes approach a situation with a machete when what you really need is a scalpel. Clients are different. Their circumstances are different. Knowing those differences takes meticulous attention to detail, experience, and relationships that have been built over time.

The organizations that get the most value from AI probably won’t be the ones that simply throw the most money at it. I think AI works better as seasoning on top of sound business practices. It can enhance what people are already capable of doing and help them do certain things faster or more effectively, but I wouldn’t want the technology replacing the human judgment that helped make the business successful in the first place.

I’d also start with clean information. Curated information. Once you fill a system with outdated material, unnecessary noise, or information you didn’t really need in the first place, going back and sorting through it becomes its own project. You can’t put the toothpaste back in the tube.

And that brings me back to where I started. There isn’t a finish line. Working with AI requires ongoing curation and maintenance, along with people who understand the business well enough to keep pointing it in the right direction. As the technology changes, that work continues, too.