Forget the Hype. Focus on Practical AI

It's been a year of ultimate AI hype: From trillion-dollar data center plans, to an unending stream of posts about how amazing AI will be, how bad AI is, how successful AI deployments are, how prone to failure AI deployments are, to the dramas going on in startups and big software companies.

The current hype environment has the odd effect of making many feel like they are both falling behind in some “AI race,” while also deeply questioning whether this technology is going to be a huge waste of everyone’s time.

The current hype environment has the odd effect of making many feel like they are both falling behind in some “AI race."

On the positive side: Unlike some other technology waves, there are about a billion people around the world using AI chatbots, and companies are investing untold billions of their own money, not just betting with investor cash.

So there clearly is something here, and many are finding value in small and large ways.

On the negative side: AI can often look great in a demo and be fiendishly difficult to make actually work reliably in practice.

What can look like an “easy button” often produces errors and misses key information.

Making it work at scale is every bit as hard as making any software work at scale. There is no easy button.

So, where does that leave us in 2026?

So, what is AI actually good for right now?

The answer is simpler than it might seem: AI can automate much of the effort of gathering, summarizing, translating, and analyzing information that is low-value-add but often a huge time waster.

For example, AI can already help with:

• Instead of hunting through reams of documents to find answers, AI can rapidly search and surface key information.
• Instead of filling out the same information over and over, AI can complete and cross-check data entry.
• Instead of reading a 700-page specification document, AI can quickly give workers the information they need for the phase they’re currently working in.

The list of these types of data- and document-based tasks that AI can help with grows by the day.

Read it: Facility condition assessments are one place AI can be a major advantage.

The real problem with AI hype

The problem with AI hype is it often confuses how technologies really progress over time.

Just because you can do something now does not mean you will automatically be able to do what you think is the obvious next step in that technology’s progression.

The clearest example of this is self-driving cars.

In 2018, everyone saw Teslas able to drive on the highway and assumed full self-driving was just around the corner. 

It’s 2026, and that problem has still not been fully solved.

In 2018, everyone saw Teslas able to drive on the highway and assumed full self-driving was just around the corner. It’s 2026, and that problem has still not been fully solved.

Those Waymos and Tesla Robo-taxis all still either have to follow a careful map, or have a human operating the vehicle remotely, or both.

We expect that to improve, but the point is that anyone in 2019 who was counting on the imminent arrival of full self-driving was in for a long wait.

What’s here today is still worth taking seriously

We have real benefits that are proven, but they remain limited.

Your job is not going to change as much, or as fast, as the big news articles would imply.

However, there are many, many small ways that AI tools can make your job easier, help you understand risk better, and get more done.

And the best part of this is, because AI is less all-powerful than some are promising, you and your team have the time to think about your work through the lens of how you can evolve what you’re doing to use these tools more, including some of the promised benefits that are not here yet.

The key is to balance skepticism that AI will be as competent as is being promised, with a recognition that this is in fact happening, and AI is already making the systems you know and trust better.

From assistants that are right in the software, to configurability that works with everyday language, to automations that save you time, AI has made its way into the software you trust, and the teams that publish that software are working hard to continually introduce new ways to make it work for you.

This is the beginning, not the end state

For those who remember the birth of the world wide web, this all seems familiar: Early promise, then some green shoots of what’s real, lots of drama, then a blooming ecosystem. That means entirely new ways of working, from apps and websites, to software that doesn’t crash, to documents that don’t need saving every five minutes, to mobile apps, and more.

AI will go through the same sort of evolution, from chatbots in 2022, to early agents in 2025w, to ... we don’t quite know yet.

You are not behind in AI adoption

Wherever AI goes next, keep in mind it might seem fast, but the real world takes time, and you are not behind.

At least not yet.

Have you heard of kBuilder Canvas? See the latest ways Kahua is embedding AI into the platform.