Got AI FOMO? 3 bold but realistic bets your business can try today

Mininyx Doodle / iStock / Getty Images

Follow ZDNET: Add us as a preferred source on Google.

ZDNET's key takeaways

• Bridging the gap between aspiration and achievement isn't easy.
• Smart business leaders focus on projects that deliver top-line growth.
• They support staff, work with peers, and encourage experimentation.

Research suggests CIOs are under pressure to roll out AI-enabled services. However, David Walmsley, chief digital and technology officer at jewelry specialist Pandora, told ZDNET it's crucial to avoid being carried away by the hype.

"The big thing for me is the amount of FOMO in this space," he said. "I know every CIO is on the receiving end of that discussion, with business leaders saying: 'Every other company is doing something radical.' But here's the thing: they're not; they're really not."

Also: 5 business leaders on how to balance innovation with risk - and turn your ideas into action

In the case of Pandora, Walmsley works with the company's senior leadership team and its vendor partners, including Salesforce, SAP, and Microsoft, to develop data foundations for exploring emerging technologies.

He reflected on his firm's nascent dabbles into AI and suggested digital leaders must focus on the things that make their business special and apply AI to those areas: "Keep your peers and leaders focused on how to apply AI to your points of strategic advantage."

Pandora focuses on "three big bets" in AI -- selling and service, product development, and back-end automation. Here's what those bets mean for the business.

1. Use agentic AI to improve customer service

Walmsley said the in-store customer experience is the gold dust -- or, as he suggested, "silver dust" -- at the heart of Pandora's processes.

IT and business professionals at Pandora must focus on how AI can be applied to create a step change in customer service, and that process means using agentic AI.

Like many other retailers, the company uses a chatbot to help its customers complete their orders.

Unlike other retailers, about 50% of that conversational traffic uses a next-generation chatbot powered by agentic AI and the Salesforce Agentforce platform.

Also: The AI complexity paradox: More productivity, more responsibilities

Walmsley said developing this capability has so far involved an eight-month partnership with Salesforce, but there's more to come -- and Walmsley said the potential is huge.

Walmsley: "Keep your peers and leaders focused on how to apply AI to your points of strategic advantage."

Pandora

"The Holy Grail for us is to then use the same agent to engage customers in sales conversations, so the agent would say something like, 'Oh, it's your daughter's 16th birthday, and she likes butterflies,'" he said.

"That's the strategic opportunity when I talk about selling and services. It's like the moonshot -- we can do service online, but selling to people via agentic AI is going to be landing the person on the moon."

Walmsley said this process involves a learning journey. Adopting agentic AI is less straightforward than many perceive, and the aim is to lower the avoidance rate, where people eschew agents in favor of chatting with a human assistant.

"Selling our products involves human emotion. Our retail experience is an active sale. It's about asking the customer, 'So, what's the person you're buying for interested in?'" he said.

"Capturing that emotion is the real challenge. We're learning from the best store associates we have around the world, in terms of how they sell. But we think AI in particular could be an interesting strand for us in terms of building that conversation."

2. Exploit models to boost product design

Walmsley said the application of AI to Pandora's product development process involves a combination of approaches.

"We already use AI in detailed back-end areas, like estimating the cost of manufacture and the time it takes to make a particular piece based on the design," he said.

The company uses probabilistic models to make forecasts based on the data from the pieces of jewelry its team has designed over the years.

These predictions are powered by a mix of bespoke and off-the-shelf models, some of which have been used for five years or more.

"Probabilistic models are important when you're trying to plan forward," he said. "We'll also use models to consider whether we think products can be made more simply and still have the same commercial impact."

Also: Forget quiet quitting - AI 'workslop' is the new office morale killer

Walmsley gave the example of Pandora using data from its two million customer reviews to hone the design process.

"My team has built a review synthesis model to pull out thematics for quality control, manufacturing, design, and so on," he said.

"It's one of those massive data assets that a lot of companies have, and my team built the model by combining three off-the-shelf third-party language tools that were then deployed discretely into an overall machine that they were building."

In short, the company exploits emerging technology across all areas of the product development process.

"There's also an area like design tools, where our design teams already play with visual design tools and the enhanced experience," he said.

"It's just thinking about the different discrete ways we could use AI across the new product development journey so that, ultimately, we're getting more of the winning products into the market quicker."

3. Automate to create a smarter organization

Walmsley said his team is also exploring how AI can help automate back-end processes. The aim is to push best practices from IT professionals to all employees.

"We're working in our backyard, in digital and technology, on tooling and automation, because it feels like there's nothing like creating a worked example that shows the rest of the business that we think this is something that can move us forward."

While AI can help boost operational efficiencies and reduce operating costs, Walmsley said the application of automation is also about creating a smarter organization.

"Raising the quality bar in terms of the working environment is also part of AI," he said. "If you can save people from sitting through long meetings where people share their opinions endlessly, you can raise the quality of people's working lives by having much more focused pieces of work."

Also: This is why your company is transforming into an autonomous machine

Walmsley described the efforts by big-name IT companies to automate back-end productivity tasks as an interesting space.

He spent some time in Redmond with Microsoft's tech experts a few months ago. Before seeing the technology in practice, he questioned the potential impact of generative AI and agentic technologies on work practices, particularly in terms of return on investment.

However, Walmsley was impressed by how Copilot Studio, an environment that helps teams build AI agents, can be used to enhance workflows. He also expects other tools from the tech giant to form part of Pandora's approach to automation.

"The folks inside Microsoft are using Copilot Studio themselves to build their stuff. I thought that was interesting," he said.

"My best guess right now is we'll probably end up using Microsoft Teams as the front end in terms of things like asking HR about how many days of vacation you've got and helping you to process invoices."