ZURU cuts product development to five months with AI

ZURU has built an AI-based trend intelligence system on Amazon Web Services to shorten product development times, cutting its cycle from 12 to 18 months to five months.

The New Zealand-founded consumer goods group said the system helped one product go from concept to shelves in five months and put it on track to generate USD $20 million in its first year.

The shift reflects pressure on toy and consumer brands to respond more quickly to social media trends, which can rise and fall within days. Traditional methods, ZURU said, left teams too slow to act while online interest was at its peak.

The system analyses large volumes of social media content across TikTok, YouTube and Meta. It processes 20,000 videos a day, according to the company, including more than 2.5 million TikTok videos this year, to identify recurring hooks, patterns and emotional signals in posts.

That information is then used to shape creator briefs in hours rather than weeks, allowing marketing and product teams to spot trends and act while demand is still growing.

Brittany Oliver, Chief Marketing Officer at ZURU Toys, described how the company tracks online behaviour.

"We used to have social teams watching social media feeds to identify trends, then taking one to two weeks to produce content against what they saw. By the time the content was live, the trend had already faded away," Oliver said. "To be globally competitive, we needed to move faster, so we created an AI system that watches content at massive scale, identifies what's working and why, and feeds that intelligence directly into our content creation pipeline. AI makes it cost effective to have a real-time pulse on cultural trends."

One example came when the system detected rising interest in stickers on social media. ZURU then developed and launched a related product in five months, which it said is on track to generate USD $20 million in first-year revenue.

Another involved its Fuggler plush toy brand. The system identified French-language TikTok creators discussing the brand without paid support, prompting ZURU to redirect content spending towards French-language creators.

Within 14 days, Fuggler recorded an 84% increase in reach and 1.6 million additional views, according to the company. ZURU presented the result as an example of how the system can shape media spending decisions as well as product development.

Experimentation model

ZURU said its approach relies on frequent testing and early rejection of weaker ideas. It reported that 50% of products and 90% of content are retired before reaching scale.

Oliver said the figures reflect a deliberate strategy rather than failure.

"We are proud of these numbers," Oliver said. "If our success rates were higher, it would mean we weren't experimenting enough. Instead, we are giving marketers a system that never stops watching, so they can accelerate what they do best."

The broader backdrop is a market in which newness plays a large role in annual sales. ZURU said 30% to 40% of yearly toy sales come from products that did not exist the year before, making speed to market a significant commercial factor.

The company also said it produces more than 40,000 pieces of content annually and has achieved 10-fold revenue growth in less than a decade. It added that a single trend-driven post generated more than USD $1 million in earned media.

Beyond toys

ZURU is also applying similar methods beyond toys through ZURU Tech, its home design and construction arm. It is developing applications intended to let customers design homes, with manufacturing handled in its own factories.

AWS provides the cloud infrastructure behind the group's trend intelligence system and wider software work. ZURU linked that infrastructure to its effort to scale testing and development across different parts of the business.

Oliver said the company sees AI as a way to help staff respond faster to customer demand.

"AI has unlocked unprecedented pace for our business," Oliver said. "Your customers are telling you what they want next. The question is whether your systems are listening - and how quickly you can do something about it."