Artificial intelligence (AI) is poised to make major waves across nearly all industries–but how can companies determine where to integrate the technology to improve business processes?

A panel of AI experts met at Mobile World Congress Americas in San Francisco on Tuesday to discuss just that. The panel included the following members:

Danny Lange, vice president of AI and machine learning, Unity Technologies

Gunnar Carlsson, co-founder and president, Ayasdi

Kris Bondi, CMO, Neura

Michael Fitzpatrick, COO, PullString

Soma Velayutham, global head of AI and deep learning for telecom, NVIDIA

Here are the six tips they offered business leaders interested in adding AI into their companies.

SEE: Special report: How to implement AI and machine learning (free PDF)

1. Accept it

Businesses need to understand the benefits AI presents, Velayutham said.

“Moving along the journey of AI is a little like addiction–you have to accept it first. If you’re in denial, it’s not going to happen,” Velayutham said. “AI is actually a process and a journey–it’s not a switch you turn on, and everything becomes at AI.”

Executives tend to be performance-driven, and want to think of AI as a box that will immediately produce results, Velayutham said. However, AI is an evolutionary technology that gets smarter and smarter as it learns. “You need to give it some time,” he added.

Within the next three to five years, AI will be so ubiquitous that it won’t seem like the futuristic tech it is currently considered, Bondi said. “It will be in so many things that we won’t need to say ‘It’s AI-enhanced.’ It will just be driving everything in the consumer world, maybe even more than in the business world,” she added.

2. Look to adjacent industries

If others in your industry have not yet adopted AI technologies, you should look to how adjacent industries are doing so, Velayutham said. “Understand what’s happening there and how they have succeeded, and bring that to your current industry,” he added.

3. Invest in talent

“AI is all about talent, and you need to create fertile ground for talent,” Velayutham said. While many developers are trained to think logically, AI requires them instead to develop algorithms that drive the logic. “Train and retrain your staff–it’s a long term journey,” he said.

4. Connect AI to your Key Performance Indicators

Even before bringing AI into the conversation, it’s important to consider the company’s Key Performance Indicators (KPIs) and where they can be improved with technology like AI, said Bondi. “If you don’t look at it from that way, it becomes a ‘nice to have’ or a shiny toy, and it’s the first thing that isn’t supported,” Bondi said. “But if you say, ‘The company’s KPIs are X, Y, and Z, and an AI approach can help us improve this much,’ then suddenly it gets buy in and will be able to move forward.”

5. Leverage the platforms available

Businesses can build upon many AI platforms today that did not exist five years ago, Velayutham said. “You need to focus on the use case and business value, and not spend too much time going after platform issues,” he added. “A lot of companies are focusing on platforms, but we should leverage those platforms and focus instead on the business value we want to drive out of them.”

For example, five years ago, running a process that would tell you what was in an image was incredibly expensive and laborious, Fitzpatrick said. Today, you can send that image to an API and get a result in a fraction of a second. “The progress in a few years is remarkable,” he said. “Accessibility to the benefits of these platforms that have been developed are increasingly being made available to individual companies and developers for free or nearly free. It will push the field forward fairly quickly.”

SEE: Big data in 2017: AI, machine learning, cloud, IoT, and more

6. Consider the people impact

At least in the near future, people and AI systems will collaborate, Carlsson said, with major implications for everything from data labeling to privacy. “In terms of installing AI or machine learning into company processes, one has to consider the people you have on the ground, who will be interacting with those processes,” he said. “It’s a huge part of what we do.”

It’s unlikely that AI will destroy many jobs, Velayutham said. “I think it will enable a lot of jobs,” he added. “We have human crises we need to solve, and AI is perfectly suited for that. People tend to take AI on the negative side, but I think it’s more positive than negative for the human race.

“The human race is extremely adaptable,” Velayutham added. “I don’t think AI is going to take over. I think we’ll all augment AI into our lifestyles, very much the way we’ve augmented mobile phones into our lifestyles. We will have a lot of AI assistance.”