A pragmatic approach to enterprise AI

2024 is over. So, too, should be the approach to AI of throwing everything at the wall to see what sticks. It’s 2025, and to get the most out of AI, businesses must be more pragmatic.Early investments yielded small successes in the market. But the juice hasn’t always been worth the squeeze. Eight in 10 AI projects fail or encounter significant setbacks or cost overruns.

There is no doubt that companies will continue to invest in AI. But in 2025, businesses will need to see a return on investment (ROI) on such efforts—and they’ll demand enterprise-level, predictable, repeatable, and scalable solutions.

Here are six ways your business can take a pragmatic approach to AI.

Start at the periphery – not the core

AI is full of promise, and generative AI has been groundbreaking. Businesses clearly appreciate the potential these technologies have to drive efficiencies and unleash new revenue streams.

But while many companies want to transform their businesses, most lack the AI expertise to truly impact their bottom line. Trying to transform the core of your business with a technology you’re not completely familiar with and that’s constantly changing can be extremely challenging.

Gain experience and expertise by first using AI at the periphery of your business. That may involve implementing AI to improve your help desk, your human resources operations or marketing efforts. This approach will provide value by allowing you to transform your business methodically, one step at a time.

While transformation will be incremental, if an issue arises, this measured approach won’t make it front-page news and the impact will be minimal. Conversely, an issue with an enterprise-wide change could shut down your business operations or worse, lead to a lawsuit because your chatbot provides incorrect information to customers.

Once you’ve exercised your AI muscle at the periphery of your business and learned from your successes (and failures), you’ll be in fighting form, trained and ready to transform the core of your business.

Set the right expectations

The potential of AI is truly limitless. But don’t expect to replace your entire workforce with AI, and don’t let others believe this either. You can’t just flip a switch with AI and you’re done. It doesn’t happen overnight.

To get where you want to go, embrace a methodical approach. Be ready to put in the work and time needed. And collaborate with the right experts to drive results. Plan, develop, implement, assess, adjust and repeat.

Identify and realize ROI

Defining AI’s ROI can be extremely difficult because it usually comes down to how it directly benefits your business.

For the best results, start by defining what success looks like. Identify the outcomes and improvements you’re aiming for. Then, pinpoint the ideal ways to measure against those goals. This might be increasing your marketing team’s ability to generate new leads, for example.

When you can demonstrate how you’ve leveraged AI to move the needle, you’ll be in a better position to invest in AI again and drive improvements across other areas of your business.

Make the right investments at the right time

Once you’ve defined the outcome you’re aiming for with AI, determine what’s needed next from a data landscape perspective to make that happen.

This might strike fear in the hearts of many. Especially those who may not know what data they have and where it is. You may be thinking your data is in no way, shape or form ready for an AI effort. But you don’t need to fix your entire data estate at once. Just focus on what’s needed to reach your outcome.

Later, you can focus on modernizing the rest of your data estate on an as-needed basis.

Leverage training and education

Gaining the greatest value from AI investments also requires proper education. Anyone at your organization who engages with AI technology must have the skills to support and maintain it. This isn’t solely an IT team requirement. It’s a two-sided equation that also includes AI end users.

Define best practices and show end users within your business how to properly use AI tools. This may be educating employees on how to properly engineer questions for AI and/or how to identify AI hallucinations when the answers they get from AI don’t seem right.

Some education may be informed by market knowledge and delivered via HR. Other training may be vendor specific.

Identify expert partners

Getting pragmatic with AI is like going to the gym. Most people who are new to the experience may not know where to start, which can be extremely overwhelming. But at a gym, you can work with a trainer to set and reach your goals, like losing weight or building muscle. You can take a similar approach to improving your business processes and company.

Because data is critical to AI success, seek a partner that provides a solid data foundation for innovation. Make sure they’re methodical in their approach to AI, aiming to simplify the complex. Pick a partner that positions you to scale, bakes in sustainability, and delivers business results.

Look beyond technology alone to select a partner with deep expertise in your sector. Relevant operational and industry experience grounds a company. It ensures they understand what’s important in your business, the processes you use, and how you want to improve upon them. This is critical because if you just throw a bunch of technology at a problem, you’re just going to be stuck on a hamster wheel—paying a lot of money without really moving your business forward.

Taking a pragmatic approach to AI, with a partner that knows both how your environment operates and the technology that AI requires—one that has transformed its own businesses in the same sector using the power of AI—will position your company for success.

Avoid stumbling in your continued AI journey and consider these six ways to approach AI. Doing so may prevent major setbacks and significant cost overruns.

Jason Hardy is the CTO for artificial intelligence at Hitachi Vantara.

—

Generative AI Insights provides a venue for technology leaders to explore and discuss the challenges and opportunities of generative artificial intelligence. The selection is wide-ranging, from technology deep dives to case studies to expert opinion, but also subjective, based on our judgment of which topics and treatments will best serve InfoWorld’s technically sophisticated audience. InfoWorld does not accept marketing collateral for publication and reserves the right to edit all contributed content. Contact doug_dineley@foundryco.com.