At Stripe Tour Sydney, the message was clear: agentic commerce is already here, and the businesses winning are the ones that rebuilt their data and pricing infrastructure first.

I attended Stripe Tour Sydney this week. The cost structure of doing business has changed, and the businesses pulling ahead are the ones who rebuilt their data and pricing infrastructure before they needed to.

Below are the key takeaways:

→ Agentic commerce isn't theoretical anymore. 81% of ANZ consumers are already using AI to research and buy. 76% of businesses say that AI-referred traffic converts higher than traditional channels. Belroy's take was blunt: none of it works without structured, unified product data first. Discoverability is a data problem before it's a distribution one.

→ Canva's Cliff Obrecht on the economics: cost-to-serve on AI features has gone from cents a month to many, many more cents a month. The free-to-paid graduation model that funded a decade of growth doesn't map cleanly onto AI products — the path to monetisation likely needs rethinking, not just re-tuning.

→ On pricing: 'billing asks did we charge correctly; monetisation asks are we capturing the value we create.' Panel consensus was that pricing needs one empowered owner, not a committee — and that most infrastructure caps growth long before nerve does.

→ Stripe's own data: business formation is at record highs, and it's overwhelmingly solo founders — 63% of Q1 Atlas incorporations were single-founder, with the 2026 cohort tracking well ahead of 2025's pace. Their framing, via Coase's theory of the firm: AI is lowering the cost of coordinating both inside and between companies, so the likely shift is toward smaller companies doing more, and transacting more with each other.

Across all four sessions the common denominator was first-party data infrastructure — it's what makes agentic distribution, pricing agility and lean scaling possible in an agentic economy.