Thomas Cornely, SVP of Product Management, Nutanix.

Tech is evolving faster than ever. If you apply the Gartner Hype Cycle to generative AI, for example, it's gone up incredibly quickly, peaked and is already on its way down to the Trough of Disillusionment. The way things are going, there's no question we will eventually get to the Plateau of Productivity—where a technology is deemed to have reached mainstream adoption.

Some enterprises have spent a lot of money investing in GenAI, while others are just gearing up to free up the budget. In all cases, there's a clear admission that this is going to be harder than we think. The mood has shifted quickly, but the story is far from over.

From Bots To Agents

I've been in Silicon Valley for more than 25 years now. We've seen several of these cycles, and they all follow the same phases: People get the spark from Silicon Valley, are shocked by the tech and wonder about all of the possible ramifications. In the meantime, people in Silicon Valley roll up their sleeves, work through the kinks, experience a ton of failures, note the areas where the tech is getting traction and eventually have a massive impact.

We saw it with the internet, Big Data and SaaS. Look beyond the return on GenAI headlines—this is happening. GenAI is like every tech revolution we've gone through.

The biggest shock is how fast it's moving through all of these phases. We've seen crazy hype almost immediately paired with hard questions (e.g., the reaction to early examples of hallucination), but in the background, there's already a lot of activity around the areas of application. GenAI has quickly moved from chatbot to productive agent to "we're going to replace everybody" to "we're going to make everybody more effective and more productive."

Some Industries Are Tougher To Crack Than Others

Take law firms, for example. In the early days of ChatGPT, friends of mine in law firms were openly wondering about the short-term impact of GenAI on the role and need for junior lawyers. ChatGPT would take care of preparing briefs, doing the research, preparing arguments and directly feeding this to senior partners. It would be faster, more efficient and more cost-effective. What followed instead were stories of lawyers running into massive legal issues due to hallucinations.

Now, some people are questioning whether industries like legal and finance are appropriate places to let the reins go and unleash GenAI due to the importance placed on being accurate. We probably have some ways to go in these verticals to make sure there's nothing that could go sideways.

There's a good reason for this sobering realization. For example, applying GenAI to banking workflows is especially difficult given how regulated that industry is. It can't simply take any application powered by any random AI foundation model. It has to know that a model is not going to veer off and create compliance issues. This is a massive hurdle to overcome; keeping "generic" tech compliant when it's effectively a black box is incredibly difficult.

There's a huge tension here between the broader GenAI applications that are going to take more time and the specific workflows and business problems where operational AI is moving at pace.

Amplifying, Not Replacing, People

It's becoming clear that it's better to focus on areas where you can tolerate more risk. For other areas, stick to complementing and making somebody more effective at their job. It's critical to always have a person involved to apply what humans offer.

We're moving from hype to reality much like when we moved from log tables to slide rules to the calculator. Operational AI is about accelerating processes and people's productivity. It's not about replacing people—it's about amplifying what people are doing.

At Nutanix, we started using GenAI to improve our support by providing chatbots to our support representatives, providing quick and easy access to information that helps guide them in the right direction. The human element continues to be important, and we're also working with our engineering teams to leverage GenAI to work faster so our teams can focus on the more complex work while GenAI solves the more straightforward tasks.

We are not unique in incorporating GenAI into our organization. I know one company that is applying it to internal data to ensure call center agents are compliant. The company records all of its customer calls and feeds them into a GenAI application that flags anything suspicious for a human to double-check.

Previously, the company's employees would cherry-pick calls, listen and try to figure out if they have to worry about a given interaction. Now, they use GenAI as a prefilter and apply their resources to what deserves a second look.

This is a prime example of using GenAI to amplify and accelerate the existing process—complementing what employees were already doing while relying on people to make the final decision.

Operational AI Is Here

GenAI is already going down to the Trough of Disillusionment because it captured our imaginations as a technology that could do anything and everything. That predictably turned out to be too much to ask. If you look more closely, however, operational AI is rising fast. Many of us were impressed by what we could do with ChatGPT and some of what followed from Meta and Google. There was a surprise factor initially, and then people and businesses wanted to know how to use the tech effectively.

Now, enterprises are applying AI to optimize processes; the shift from bots to agents is ramping up quickly. This could not have happened without the GenAI big bang moment.

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?