Based in Hong Kong, Prenetics is a NASDAQ-listed, health science company. Its flagship brand is IM8: a range of nutritional supplements co-founded with David Beckham, backed in an equity deal by US football club Inter Miami – current home of Lionel Messi – and represented by brand ambassador Aryna Sabalenka.

From its launch in December 2024 up to April this year, the IM8 brand has recorded sales worth $186m, achieved without any addition of headcount. “IM8 products are 100% direct to consumer and we ship to 42 countries,” Prenetics CFO Stephen Lo explains. “With 40% of our business, the US is our biggest market, followed by Canada, then the UK and then continental Europe.”

Importantly, Lo notes, Prenetics runs IM8 as “a 24/7 machine”, with the brand’s commercial impact in its various markets assessed and reassessed around the clock. Given the time-zone differences, Lo says: “That’s difficult for our finance team in Hong Kong to manage without artificial intelligence (AI). So, we use AI extensively, for tasks such as customer acquisition, reviewing marketing performance and monitoring operational KPIs.”

The AI is connected to the core, ecommerce engine behind Prenetics’s online store, and Lo receives granular updates from the software two or three times a day. In practical terms, he says, AI has become a powerful ally in the company’s growth story.

AI is used to speed up financial reports

Lo says that under the typical financial closure process, it would take months to build an IFRS-compliant financial statement that Prenetics could use to inform its business strategy. But for a dynamic and ambitious multinational startup brand such as IM8 – which is already earning at least $10m per month in revenue, from a standing start of $600,000 in December 2024 – relying on that standard timeframe is not a viable option.

“It’s far too long a feedback loop,” Lo says. “We want to know what’s happening in real time. That means looking at our marketing campaigns in individual territories and evaluating what works and doesn’t.

“For example, maybe we should push our links with David Beckham more in certain countries, or focus on Sabalenka as our brand ambassador in places where tennis is growing more popular. Perhaps we should adjust how we’re engaging with football and tennis players on social media. And perhaps that requires lots of A/B testing to see how we can get the message across most effectively.”

All of this is extremely difficult and simply not scalable at the rate the business wants to grow, Lo stresses, if carried out by humans alone.

AI allows for real-time, detailed data analysis

To that end, the company’s AI is plugged into every stage of IM8’s product and marketing lifecycle. Alongside promotion, it scrutinises key areas such as revenue generation and the effectiveness of the brand’s third-party logistics provider, all the way up to accounts closure. That thorough coverage enables Prenetics to detect anomalies (such as problematic orders), investigate why those issues have occurred and remedy them in a prompt fashion.

Lo describes the quality and detail of the information the AI provides as “amazing,” and says that the software has freed him and his team to sharpen their focus on troubleshooting and problem solving.

As well as applying the AI tool to IM8’s product strategy, Prenetics has enlisted its support for employees’ admin, enhancing staff productivity. Indeed, Lo says, the AI sits within his own mailbox, calendar and desktop financial software as an on-tap assistant to speed up his completion of routine tasks. “This usage operates strictly within corporate-approved platforms,” he notes.

The importance of training – for AI and staff

Following on from that, Lo highlights a critical risk-management point: that integrating an AI with your employees’ everyday systems requires supervision, guardrails and monitoring.

“AI is context dependent and can hallucinate if not properly trained,” he says. “So, before you can use it for work tasks on a day-to-day basis, it must be exposed to your digital office environment, including your mailbox and so on, to pick up your writing style and gather information that will help it understand your thought processes.”

Alongside that need for AI to familiarise itself with its users, staff training is essential for preparing employees to make the most of the technology in the effort to be as productive and efficient as possible. Meanwhile, a ‘human in the loop’ step is vital for checking staff AI outputs before they go out to business partners.

AI policies must come from the top

Looking at risk on a broader level, Lo says it’s very important to have an AI policy in place. “This is not a typical tech policy where you delegate the risk management to the head of IT. It has to come from the top down,” says Lo. “Everyone in the C-suite must understand the risks, potential and opportunities that stem from using AI and decide how best the business should navigate them.

“We use third-party providers to manage our AI risk – particularly in relation to security of customer data and maintaining compliance with GDPR and other regulatory safeguards.”

Use a sandbox to test use cases

For developing novel AI tools that could be introduced into the company’s global operations, Prenetics runs a ‘sandbox’ system. This enables staff to experiment with a range of different AI models and test-run new agents or other potential solutions before they are brought into live service. “It’s a safe, secure and ring-fenced environment,” Lo notes.

Lo acknowledges the brewing AI backlash and what it means for the technology’s acceptance in the near and long term.  “The reality is that not everyone will be able to catch up with AI, because of how incredibly fast it’s evolving,” he says. “But if people are fearful of AI, then adoption is doomed to failure.

“That leaves executives with a big responsibility to build a strong narrative around their AI strategy that employees can understand, including information on how teams will be repositioned to ensure no one is left out.”