boy jobs”, as he puts it, was walking door to door selling domain names and one-page websites to local accountants, dentists and retailers. “To see how far we’ve come from static Yellow Pages print ads and one-page websites to what we’re talking about today – the progress is just unbelievable,” he told Inside Retail. That long view matters, because the constraints on small advertisers haven’t really changed: Video has always been the holy grail, but it has also been expensive, slow and intimidatingly technical. For most Australian and New Zealand SMEs, Pang noted, video has been available only to the “biggest brands in the world”, requiring deep pockets and specialist skills. The promise embedded in Amazon’s new Video Generator tool is not a shiny new format, but a way to strip out those barriers at scale. A free tool with industry‑level ambition Video Generator, now available in Australia, via Amazon’s creative studio, turns product images, existing videos or even a product detail page into six high-motion video ads in a matter of minutes. It stitches together multi-scene stories with smooth transitions, text animations and music, using Amazon’s audience insights and product metadata to decide what to highlight. Advertisers can upload social content or tutorials, and the tool summarises key clips into ad-ready formats, then tweaks logos and headlines to stay on brand. Time, not templates, is the killer feature For Pang, the real currency here is time. Amazon-commissioned research found Australian small-business marketing leaders believe AI tools could give them back roughly 7.3 hours a week – about 45 working days a year – if used to create and manage campaigns. “If I told you I could give you something that would give you back 45 days, I mean, you’d literally fall over yourself,” Pang said. In his telling, the tool’s real job is to lower the cost of entry and automate the grunt work – cutting, reformatting and multiplying creative variations – so marketers can spend more of their energy on strategy and storytelling. He is explicit that “human creativity is absolutely at the core” of what Amazon is trying to build, and that the tool is not seeking to replace “big, beautiful storytelling” but to clear the way for it. Guardrails against generic AI creative Pang is acutely aware of a growing unease that AI-generated creative will flatten brands into the same handful of visual tropes. He pushes back on that idea by positioning the generator as a first rung on a ladder of sophistication, rather than a one-size-fits-all solution. For first-time video advertisers, it opens the door; over time, more advanced capabilities within the creative studio – including a “creative agent” already live in some markets – are designed to support more complex, bespoke work. He described that creative agent as a kind of production concierge that can help advertisers move from simple, AI-assisted assets to sophisticated campaigns that combine studio-shot footage and machine-generated elements. “I think the future is agentic,” Pang stated, predicting that agents will not only speed up asset creation but also automate the loop from testing to measurement to optimisation. The aim, again, is to give marketers back the hours they need for “storyboarding and drawing pictures and bringing different content assets together, trialling and testing”. Getting closer to one‑to‑one ads Amazon, though, is hardly the first to market with AI video generation for marketers. Impressive rivals such as Runway, Sora Synthesia or Luma Dream Machine excel at producing photo-realistic clips from prompts or even detailed scripts. Across the market, the strongest AI video generators distinguish themselves by depth of reactive control and production-grade polish rather than “text in, clip out” convenience. Standout tools like Google Veo and Kling compete on capabilities such as advanced camera controls, keyframing, multi-shot consistency and 4K, HDR-ready output. Meanwhile others, like HeyGen, layer in extras like custom avatars, video translation and lip-sync. Amazon believes it has the edge because its tool can draw from product pages, customer reviews and on-site behaviour. A whisker away, in other words, from true one-to-one marketing. Pang talks about the “magic sauce” of letting performance signals flow back into tools like Video Generator to guide what gets created next. He framed it as a full-funnel system. A small brand like Sneaker Laundry can make its first-ever video, see how that asset influences consideration and purchase, and then let the platform serve different messages to bring the customer back. Over a three-to-five-year horizon, Pang expects agentic systems to automate more of that loop, so that, “You need to do it once, and your job as the manufacturer is you make the product and the technology will look after all of the rest of it.” It’s a vision that nudges the industry closer to that long-held dream: The right ad for the right person at precisely the right time, built and optimised in the background. “I’ve been chasing that,” he admitted when the conversation turns to one-to-one advertising, but with tools like Video Generator on the market, he sounds more confident than ever that the dream is finally entering striking distance. Further reading: In 2026, good marketing can’t cover up a bad product