Two Australian marketing leaders have gone on record backing generative AI in creative production, arguing the results already justify the spend — provided teams treat the technology as a test-and-learn instrument rather than a replacement for craft.

RAA's Chief Marketing Officer and T2's CEO laid out their positions in a Mi-3 interview, with RAA's CMO describing the approach as "pushing the AI snowball off the hill" — a deliberate effort to get momentum behind AI adoption now so the organisation compounds the learning early.

The case they're making

The two executives anchored their argument on results, not roadmap slides. RAA's marketing chief pointed to attribution performance, telling the publication the brand has recorded its best-ever attribution scores since leaning into AI-assisted creative work. The insurer has also used the tooling to outperform competitors in testing, according to the report.

T2's CEO framed the payoff differently: cost. For a tea retailer running continuous campaign cycles, AI-generated creative material cuts production spend enough to widen the funnel of concepts a brand can afford to test before committing budget to finished assets.

That economics-first framing matters for operators. The pitch here is not that AI produces better advertising by default. It is that AI makes iteration cheap enough to let data — not instinct — pick the winners.

Test-and-learn over replacement

Both executives stressed a test-and-learn mindset. The workflow they describe runs AI-generated variants through live market testing, with performance data deciding what scales. Creative teams shift from producing every asset by hand to directing, curating and refining machine output.

This matches what most large advertisers experimenting with generative AI now report: volume goes up, unit cost goes down, and the discipline moves to evaluation. Attribution measurement becomes the referee.

The attribution point deserves weight. Brands have long struggled to connect creative choices to commercial outcomes. If AI-enabled variant testing genuinely sharpens that signal — as RAA claims its scores show — that is a measurement story as much as a production story.

What the executives did not claim

Neither leader argued AI creative matches top-tier human craft across the board, and neither framed the technology as a headcount play. The stated value sits in speed, cost and measurable iteration. That restraint is notable in a category where vendor claims routinely outrun delivered results.

Cost came up repeatedly. For T2, a CEO-level endorsement of AI creative rests on the arithmetic: cheaper asset production funds more tests, more tests surface better-performing creative, and better-performing creative improves return on ad spend. The loop, not any single output, is the product.

Why operators should care

Australian marketing leaders rarely attach their names to AI creative endorsements this early in the cycle. RAA is one of the country's largest member-based insurers, and T2 runs a national retail footprint. When brands of that scale report best-ever attribution numbers tied to AI-assisted workflows, mid-market operators will face pressure to run the same experiments — or explain to boards why they are not.

The snowball metaphor cuts both ways, of course. Momentum is easy to claim and hard to audit. What RAA and T2 have offered so far is directional evidence: better attribution scores, competitive wins in testing, lower production costs, and a management mindset built on continuous experimentation.

Operators weighing their own AI creative budgets should ask the same three questions these executives answered: does it beat what we ran before, does the measurement confirm it, and does the unit economics hold at volume? On those terms, RAA and T2 say yes.