What a Dog Cancer Vaccine Startup Teaches Malaysian SMEs

by

Your Next AI Lesson Might Come From a Dog Named Rosie

Somewhere on your social media feed this year, you probably saw the story of Rosie, a Staffordshire bull terrier-Shar Pei mix whose owner claimed an AI-designed mRNA vaccine shrank her tumours. It is the kind of headline that makes you stop scrolling. But here is why you should care: that story has now spawned a startup called Gamgee, backed by the storied tech incubator Y Combinator (source). For a Malaysian SME owner running a business with 1 to 50 staff, whether you manage a restaurant in Petaling Jaya, a logistics outfit in Johor Bahru, or a boutique in Penang, this is not a story about dogs, vaccines, or biotech. It is a story about how hype is built, how credibility is won or lost, and what AI actually does in the real world.

What Happened

Paul Conyngham, an Australian tech entrepreneur who stresses on Gamgee’s website that he is “not a biologist,” said he used ChatGPT, Grok, and other AI tools to design a personalised mRNA cancer vaccine for his dog. He said several of her tumours shrank (source). The story went viral. Now he is launching Gamgee, which offers “personalised mRNA cancer vaccines for dogs” and, eventually, hopes to treat humans. The company is backed by Y Combinator, is recruiting for its first clinical trials in Australia, and touts a roughly 75 per cent tumour reduction in its “first documented case” (source).

But The Verge’s earlier reporting pointed out that the science is messier than the story suggests. The exact role AI played in the endeavor wasn’t clear, and researchers said the broad claims gave the technology too much credit while minimising the role of the humans involved. More importantly, the protocol Rosie followed made it impossible to say whether the mRNA vaccine was responsible for her improvement, since it was administered alongside another therapy (source). In other words: the outcome is real, but the cause is unproven.

Why This Matters for Malaysian SMEs

Think about the last time you saw a post claiming “AI did this” — a marketing video that wrote itself, a sales email that landed a big client, a dashboard that predicted everything. As a Malaysian SME owner, you are bombarded with offers for AI tools that promise to transform your operations. Rosie’s story shows you a pattern: a dramatic claim, a founder who is not an established expert in the field, and a product built on a single compelling case. That does not mean the claim is false. It means you should ask questions before you hand over your data or change your workflow. When an AI vendor tells you their tool will raise your conversion rate or cut your response time, ask for the context: what else was running at the same time? How many customers were involved? What does “improvement” actually measure?

On the other side of the coin, there is a constructive lesson. Conyngham did not wait for permission, a degree, or a large team. He took a real problem, used widely available AI tools, and built a prototype. For your business, that is the practical move. You do not need to adopt AI wholesale. Pick one annoying task — replying to the same WhatsApp enquiries every morning, summarising daily orders, drafting quotes, or generating product descriptions for your website — and test a tool you already have access to on it this week. Measure whether it saves you time. That is how AI becomes genuinely useful for a Malaysian SME: not in a dramatic, headline-grabbing way, but quietly, one workflow at a time. A viral story gets you attention; a repeatable process gets you reliability.

The Bigger Picture

What Gamgee is doing is actually an emerging category: vertical, AI-specialised services that promise speed and accessibility. The vet-facing pitch is telling: “less than 20 minutes of a veterinarian’s time, no new clinical workflows, and a finished vaccine in hand in around four weeks” (source). Even though the company cautions the timeline is based on a single case and may vary, the pitch is about convenience and speed, not science. That is the same promise many software vendors make to Malaysian SMEs — minimal effort, maximum magic. The difference between a trustworthy vendor and a hype merchant is usually in what they admit they do not know. Gamgee’s own caution is the most honest sentence in its entire pitch.

“If a business cannot clearly attribute an outcome to the process it sells, every testimonial becomes just a story. Stories attract attention; evidence keeps customers.”

The bigger pattern is that AI hype cycles are predictable. A striking claim emerges, media repeats it, a startup forms, and then reality sets in. Remember this when you see a competitor boast about their AI-powered systems. You are not competing on the boldness of your claims. You are competing on reliability, on trust, and on the experience you deliver to your customers in Malaysia, day after day. And when your own business improves because you quietly automated a boring task, you do not need a viral story. You need the result.

What You Can Actually Do This Week

  • Audit one repetitive task in your business and identify exactly where the time goes.
  • Take a single case — one product, one campaign, one customer journey — and apply an AI tool to it. Do not automate everything at once.
  • Document the result honestly. What did you change, what else was happening at the same time, and would you attribute the improvement to the tool or to other factors?
  • When an AI vendor claims a big result, ask for the context behind the number, not just the number itself.

Rosie the dog is adorable, and her story made a good headline. But the reason Gamgee is being watched closely is exactly because the evidence is thin while the claims are large (source). Your Malaysian SME does not need a miracle. It needs a reliable improvement. Start there.

Ready to Streamline Your Operations?

Technology moves fast. Your operations should keep up. AutoRunBiz builds AI systems that run your daily workflows — from WhatsApp order capture to accounting. Book a free 15-min ops audit →