In the dynamic technology landscape, where innovation is the key to success, leveraging generative artificial intelligence (AI) is an imperative for companies looking to stay ahead. That said, companies must be thoughtful about how they use AI, else they risk embarking on an expensive endeavor with little to show for it. As we look ahead to 2024, here are some tips that IT leaders should consider when investigating AI solutions.

Data is Fuel

Generative AI derives its power based on the quality of the data it pulls from. To be useful, data needs to be available and fresh. Too often, a company’s data resides in silos, with no practical way to pull it together in a timely manner. Taking the time to break down those silos and integrate your data will pay big dividends.

For example, our contactless payments brand in Brazil, Sem Parar, created an AI customer service chatbot called “Super Carol” to process inquiries quickly and accurately. That required the creation of an API that gathers data from multiple sources into one place, allowing customers to interact with the chatbot about their specific account. The result was a 5 percent increase in “contact containment,” meaning the customer stays with the chatbot until they get satisfaction instead of bailing to an agent. Super Carol also learned from its interactions, improving to 95 percent correct answers, including a better understanding of slang, local expressions, abbreviations, and spelling.

Computation Costs

Building an AI solution in-house and hosting your own models requires a large capital investment in hardware, because AI is very computationally intensive.

Most companies should consider taking advantage of existing cloud-based AI resources such as Microsoft or Amazon. Even those are not inexpensive, so IT leaders will want to focus their AI efforts on use cases where there is clear business value. They should be ready to demonstrate the cost-to-business-value ratio.

An example of a specific business value provided by AI is our fleet maintenance platform “1link Service Network” deployed by epyx brand in the UK. An experimental pilot found AI can help predict almost half of fleet breakdowns before they happen.

That steers drivers to the nearest shop for preventive service, saving the company the cost of expensive repairs. That is a clear payoff for adopting AI.

They Will Build It (So You Don’t Have To)

A best practice in IT is “don’t build what you don’t have to.” AI is in the early phase of a Gartner hype cycle, a sort of land grab as vendors rush to integrate AI into their products. Companies using these products should stick to their core business strategy rather than spending money building AI functions that will be cheaper when the vendors add them.

For example, the software development platform GitHub collaborated with Microsoft and OpenAI to create GitHub Copilot. Our software developers now use it to generate intelligent coding suggestions, blocks of code that the developers can then edit to make better. There is already lots of hard evidence that this investment is paying off in higher developer productivity. A recent McKinsey study found software developers can double their speed using generative AI.

Be Ready to Experiment

When IT leaders explore ways to use AI, they will need to keep an open mind. Experimenting with AI may produce unpredictable results because generative AI is a bit of a black box. You put in a prompt, it does its thing, and it gives you an answer. You may love the answer, you may need to tweak it, or it may not work at all. That’s different from traditional IT development where you can continue hammering on a program until you get the result you want. With AI, you may instead reach a point of diminishing returns, forcing you to cut your losses.

It’s important to remember what generative AI is and is not. It’s not a calculation engine (it is notoriously bad at basic math). It’s not a rule engine either. It won’t replace the tools you’re already using that function well. AI is simply a new tool in your IT toolbelt. It can complement the tools you’ve been using all along. Understanding that will help IT professionals reap the greatest benefits that AI can provide.