When Pick n Pay switched on Penny, its conversational AI shopping assistant, last week, the launch story was told by the retailer and Google. Less visible was a small South African software company that has spent three years inside Pick n Pay’s online operation – and in whose regular technology sessions the idea for Penny first took shape.
Catalyst Software, founded in 2016 and known as MAD until a 2022 rebrand, has worked alongside Pick n Pay’s internal teams to reimagine and rebuild the asap! apps now in the app stores, and continues to maintain and extend them, MD Iain Mackenzie told TechCentral in e-mailed responses to questions and in a recent video interview.
Penny, he said, emerged from that relationship. “We have a very hands-on relationship with Pick n Pay, and in one of our regular technology ‘show and tells’ with Enrico, actually, the idea of Penny evolved,” Mackenzie said, referring to Enrico Ferigolli, Pick n Pay’s retail executive for omnichannel. “We brought the technology implementation, Pick n Pay brought the product fit and direction, and Google provided the base LLM intelligence.”
The 2025 rebuild of asap!’s front and back ends – which Ferigolli has credited with letting Pick n Pay ship AI features faster – is the platform work Catalyst was engaged in. The retailer reported online sales growth of 48.7% for the year to 2 March 2025 on the back of the new platform.
Catalyst has grown from its four founders – Mackenzie, Jonathan Turck, Chris Kilpin and Kishyr Ramdial – to a team of about 35, distributed across South Africa, Portugal, Canada and the UK. The UK operation is headed by MD Nikita Turck. The company describes its model as “forward deployed”: senior engineers embedded inside client teams rather than contracted at arm’s length, an approach it argues suits AI work, where the gap between a demo and a production system is where most projects die.
AI agents
Beyond Pick n Pay, its current work includes modernising the internal platforms of a South African merchandising company, with some work in financial services.
Its focus is deliberately narrow: large-scale, high-volume retail in South Africa, a definition it stretches to include retail banking and digital goods. “It is where our experience and expertise over the last decade sit,” Mackenzie said. “We work within all parts of that operational lifecycle, starting with the customer-facing front ends and extending into any area that ultimately brings a better experience to the end customer.”
Watch | Pick n Pay’s Enrico Ferigolli on Penny, the AI that shops for you
The company practices what it sells. AI agents are wired into Catalyst’s own operations – including Anthropic’s Claude running in the company’s Slack, responding to requests fired at it in chat – and the team tracks an internal leaderboard of who is consuming the most AI tokens, a light-hearted proxy for who is leaning hardest on the tools.
Alongside its custom development work, the company has productised a set of AI tools it calls Catalyst Retail Intelligence, spanning image analysis, text recognition, support and automation – a sign that, like its most famous project, the firm is betting its next decade on AI in retail. – © 2026 NewsCentral Media
- Subscribe to TechCentral’s daily newsletter
- Get breaking news alerts on WhatsApp
