5 days ago
14
There is little debate that AI will revolutionize working practices, but there is less agreement about the best way to exploit this transformation.
While 90% of CIOs are piloting AI or investing in small or large-scale developments, over two-thirds (67%) haven't seen measurable ROI, according to the recently released Nash Squared/Harvey Nash Digital Leadership Report.
Also: 4 ways to scale generative AI experiments into production services
"Leaders know the technology, but they're struggling with its application in the business to create value," Nash Squared CIO Ankur Anand told ZDNET during a conversation about the key points emerging from the leadership survey.
So, how can business leaders overcome this struggle? Four business leaders provide their best-practice tips for using AI to solve big business problems.
1. Create a top 10 list
Joe Depa, EY's global chief innovation officer, said your use cases should align with your highest-value business priorities.
He told ZDNET that this alignment must be a continual work in progress. Business leaders should keep refreshing their approach to focus on the areas that matter.
"I often use a top 10 list, just to keep it simple," he said. "Here are the top 10 use cases that we'll focus on. Anything more than 10 and there's a danger people lose interest."
Depa said reviewing priorities for the top 10 list with other senior executives requires a firm, strategic hand.
"If people want to add things, I'll say, 'What will we take off the list?' Because when you add something to the list, you've got to take something off," he said. "That approach helps keep people focused. Once you have that regular cadence, updated and refreshed, people can start thinking about applying AI to use cases."
Depa said this careful strategy helps businesses avoid spending money on AI solutions that don't hit their ROI metrics.
Also: 5 ways to turn AI's time-saving magic into your productivity superpower
"It's usually when you don't have a clear use case or business value attached, but you have a cool problem that you want to solve. Those are the ones that you have to say, 'Hold on a second. I know that's a cool problem, but what's the business case?'" he said.
"That's where you can get in a bit of a rut if you go down a path of just trying to solve some problems with AI, without having a clear business case for its application."
2. Run hackathon sessions
Adobe CIO Cindy Stoddard said her IT team has used AI in many areas and works with the rest of the business to identify other use cases.
Her team used AI to explore past IT requirements and create recommendations so business analysts and product managers know what will likely be required when users demand a new service.
Also: Is your business AI-ready? 5 ways to avoid falling behind
The IT team also used AI in testing to create scripts that can be reused to automate repetitive processes.
For new use cases, the team runs hackathons that help surface applications for emerging technology within IT and across the business.
"People submit different ideas about what they think could change," she told ZDNET. "We encourage everyone to submit areas for improvement around what they see at the ground level."
Stoddard's team then works with business peers and external partners to select the best projects.
Also: The top 20 AI tools of 2025 - and the No. 1 thing to remember when you use them
"People who submit ideas put together teams. We'll also bring in some of our key vendor partners to help train staff on the technologies," she said. "Then we'll go through a development and judging process to see if the ideas deliver value. Many of the ideas we deliver end up in our production systems."
3. Learn through failure
Caroline Carruthers, CEO at consultant Carruthers and Jackson, told ZDNET about five ways to prepare an organization for an AI transformation.
However, she also mentioned something crucial to identifying the right use cases -- embracing innovation.
Carruthers said there's lots of emerging technology to test, from large language models to causal AI.
"You don't know how this tech will fit in your organization. Waiting until things are perfect with AI won't work," she said. "You need to experiment. You need to carve off a little safe sandbox, something you can start and play with to understand how your organization can get the best from AI."
Also: 5 ways to manage your team more effectively in the AI-enabled enterprise
Like other business leaders, Carruthers said it's important that the projects you support focus on the right areas. That targeting can mean learning by failure, as long as it's not too expensive.
"If you're genuinely going to experiment with AI, then you almost need to get into celebrating failure," she said. "Conducting an experiment and finding out the experiment didn't work, but learning something new, is a valid use of an organization's time. Just don't spend a lot of money doing it."
Carruthers said every AI initiative should be part of a larger project to overcome big organizational challenges.
"Innovation is about making AI part of the business, but it's doing it small, iteratively, and safely. When we talk about experiments, and especially when we talk about data, because it can solve some of the world's biggest problems, our brains tend to go, 'Oh, look at all the stuff we can do,'" she said.
"But if we try to tackle that problem, it's big, unwieldy, and we'll get bored, and we won't deliver things in time. Whereas, if you solve a small problem, it's like, 'Oh, that's nice,' and then you solve other problems."
4. Educate your employees
Tobias Sammereyer, team lead for performance engineering at XXXLutz, said many people are lulled into a false sense of security, thinking that easy-to-use tools like ChatGPT can be applied to any business case.
"We need to educate our guys how to use AI stuff properly, and how to be precise with their prompts to get what they want," he said.
Sammereyer told ZDNET that business and digital leaders must help their people understand AI's benefits and limitations before they apply the technology to use cases.
"Try to tell them what's possible, but also what's not possible, because there are two kinds of people -- one thinks AI is just hype and the other believes they can do anything with it. And both are incorrect."
Also: AI agents aren't just assistants: How they're changing the future of work today
Sammereyer said the key to success is finding the middle ground.
"Educate people, and then you can use AI, especially with generative AI," he said. "Just be aware that AI can make mistakes in the same way that a human being can make mistakes. Double check the results, and then you're good to go."
He said this process relies on your AI systems being fed enough reliable data.
"AI is like a people pleaser. The tools want to give you a good answer -- if it's correct or not, that's up to you to double-check," he said. "So, you need to be critical in your thinking and see if the AI system has enough data and is equipped to give you the right answer. Just remember that it will give you an answer, but not necessarily the correct one."
Want more stories about AI? Sign up for Innovation, our weekly newsletter.
English (US) ·