SHIFT > Local AI startup targets retail sector

By Louise Burgers, Retailing Africa Editor. New platform Productify.ai aims to disrupt ecommerce content production by using artificial intelligence. Launched by co-founders Dylan van der Merwe, owner of Happen Software, and communications consultant Andrew Trench, Productify.ai is creating compelling product content at speed and affordably for tens of thousands of products.

The new platform is powered by the kind of generative AI technology made famous by ChatGPT, and specialised algorithms which are designed to solve the ecommerce content problem. How it works, is that by using as little as the product’s name, Productify.ai can return quality marketing descriptions, features, attributes and benefits, SEO keywords, and other essential marketing content in seconds.

In an exclusive interview with Retailing Africa, Trench explains that the platform makes it easy for anyone, no matter their catalogue size, to get market-fit fast. “When Covid hit South Africa, ecommerce exploded. I observed how retailers and brands scrambled to get their product catalogues ecommerce fit. Gone are the days when a simple product name and image suffices for a print advertising leaflet. To succeed now in online retail you need comprehensive marketing descriptions, features attributes and benefits, SEO-rich material and so on, just to enter the market properly. All of this content, until now, takes a human to create and it is slow and expensive. Imagine you are a retailer starting at ground zero with 30,000 or more ranged items? By the time you get into the market with proper content, you are starting the cycle all over again – and you might be six months behind your competitor.”

He says that Productify.ai, using generative AI technology, can turn around about 20,000 product descriptions in a day, for example. The affordable price point means marketers or ecommerce leads can afford to play with content, do A/B testing for example, and see what really works to convert. Basically, it can handle the trickiest of products and content outputs designed to convert into sales. Effective ecommerce content can be generated for under one rand a product, in many cases.

Other platform features include the ability to rewrite existing content using AI and solving the vexing problem of extracting structured data, like nutritional information from product packaging, which is increasingly required for customers with special dietary or other requirements.

Processing data at scale

Productify.ai provides an API to integrate into enterprise workflows and systems and facilitates bulk processing on the platform for users to process data at scale. Content for single products is easily created too. The platform soft-launched in early April and has already secured the support of a market-leading provider of brand content, and has also been retained by a top retailer to process several thousand products for an ecommerce-enabled site launch – a job which would have taken weeks using traditional solutions – was accomplished in hours.

“As ecommerce continues to grow exponentially, we believe Productify.ai will become an essential tool for large and smaller retailers and brands, which are under constant pressure to differentiate their product ranges with compelling and accurate content,” said Trench, a former newspaper editor and media executive who also has experience in the ecommerce content space.

Trench has a new retail ecommerce launch to use as an example: “Our real-world example involves a retailer with an imminent launch of a new ecommerce play and many thousands of products which suffer from a paucity of usable content. This is a pretty usual situation for many retailers because their systems and database technologies are orientated towards brick-and-mortar businesses – not the content-rich world of ecommerce. We are able to take something as simple as the product name and optionally (although ideally), a few product attributes, and with that we can return a rich product description.

“In fact, it takes about 20 seconds for us to provide three different versions of a marketing description which run to a couple of hundred words each, generate a short description, a marketing headline, several dozen SEO keywords and a categorisation. A skilled person would likely take about half an hour to do the same work. But we are also careful to stress – in fact this is explicit in our terms and conditions – that all outputted content needs to be human-reviewed before distribution. We are, after all, dealing with an AI model here, not an intelligent being despite the hype. So, a human-in-the-loop approach is critical at this stage until the AI develops further,” reiterates Trench.

Real-world retail solutions

The proudly South African-made platform has attracted interest from users across the globe from countries as diverse as the United States, Australia, Ireland, Germany, the UK, Brazil, Kenya, Uganda, and Nigeria among many others. “It is rewarding to see Productify.ai create content for real-world products on a daily basis and while we are thrilled that the platform is delivering as we intended, we continue to work collaboratively with our users to develop new features and capabilities to constantly expand the power of the platform,” added Van Der Merwe.

Van Der Merwe’s software business, Happen Software, has created robust specialist enterprise systems for leading companies over the past decade. Trench and Van Der Merwe have collaborated on other projects, including an award-winning election results platform for TimesLIVE and News24’s CrimeCheck application, among others. “We see AI as a huge opportunity for entrepreneurs from Africa to create products and services that reach beyond boundaries to solve problems in the world,” says Trench.

As Trench adds, AI can do much of the heavy-lifting of basic marketing content and it is increasingly used in other spheres of the industry like in personalisation on ecommerce platforms or in managing logistics or fulfillment. It is only a matter of time before it becomes ubiquitous in this space too.

“AI is a force multiplier. In retail where margins are razor thin, business leaders know that spend needs to be kept focused and used intelligently in every aspect of their businesses. It does not make sense to use armies of humans to do the work which an AI can do competently. Sure, keep skills to ensure quality control, for example, but otherwise direct spend to where it will differentiate you in the market, fight for market share or do the other things that are important for retailers.”

SHIFT> is a new interview section to spotlight sustainable brands and entrepreneurs creating innovative retail solutions.

Louise Burgers is the Publisher, Editor and Co-Founder of RetailingAfrica.com. She has spent over 20 years writing about the FMCG retailing, marketing, media and advertising industry in South Africa and on the African continent. She also lectures post-grad students in Marketing and Advertising Communications at the Red & Yellow Creative School of Business; and works with the global Chief Marketing Officer (CMO) Council in the Africa region on editorial strategy. Specialising in local and Africa consumer trends, Louise is a passionate Afro-optimist who believes it is Africa’s time to rise again and that the Africa Continental Free Trade Agreement (AfCFTA) will be a global gamechanger this decade.

Receive the Retailing Africa newsletter every week • Subscribe here