Brad Audet describes AI as already embedded across the marketing toolset, in media mix modeling, media scenario planning and audience design, whether or not those tools are labeled as AI.
His framing throughout is augmentation rather than replacement. He does not expect AI to replace humans or human intellect, and describes the ambition as using it to unlock how the human brain can work differently.
Key findings
- The starting question is a problem, not a technology. The company has not asked where AI can be applied so much as what problems need solving and how to address them most efficiently.
- The agency platform in use has two halves. One holds brand and audience models that support synthetic audiences and hypothesis evaluation, and analyzing ten years of brand data has quickly exposed gaps that would otherwise have taken far longer to find.
- The other half is execution: asset creation, audience development and media distribution, with the intelligence applied quickly into a test and learn environment and an automatic optimization loop between the two.
- Where he believes the company is genuinely ahead of its industry is data sharing with retailers. A bilateral data sharing agreement supports a 360 degree view of the customer, with a partitioned customer data platform for every dealer collecting first party data.
- His view of the category is less flattering. As marketers, he says, the automotive category is a bit behind everybody else.
Where he expects AI to matter next
The first area is customer experience. AI can help address complaints more quickly and thoroughly the first time, and can reduce labor cost by prioritizing or scoring incoming issues so they are routed more effectively.
The greater value, in his view, is upstream. Aggregating very large volumes of data to find where the company can solve problems and explain things to customers earlier and more meaningfully.
The second area is the vehicle itself. A connected car generates tens of terabytes of data continuously, which would be very difficult to assess for customer behavior opportunities without AI.
Looking a decade out, he suggests the car could become a biometric device comparable to a wearable, potentially reading blood pressure through a camera in the mirror.
He is more cautious about the near term. Autonomous driving depends on extraordinary computing power and remains unsettled, and he expects significant innovation in the category to slow for a while.
Why brand matters more, not less
Asked about the worry that shared tools produce uniform creative, he takes the opposite position: brand is more important in this era than at any point before it.
His reasoning is mechanical. AI defaults to reading masses of information and producing a prediction for the customer, and the only thing that can overcome that prediction is a decision bias or brand bias.
The practical instruction he gives his teams is to create a moment of pause. When a competitor is recommended in a search, the brand’s job is to make the customer question whether that recommendation is right for them.
He connects that to humanized growth directly: understanding the relationship between the consumer and the output of the AI, and what role brands can play in intercepting and guiding it.
The company’s own purpose is the anchor. Enriching lives means taking care of people and moving them forward, which he says cannot come from artificial intelligence or algorithms because it is people helping people.
The leadership position
He describes a live tension inside an enterprise technology transformation. On one side are people leading it who are technology centric and treat the technology as the hero. On the other is his own insistence that the benefit has to be the value to the human.
His argument for that is customer-facing: nobody cares what the platform behind an experience is, and anticipating a problem before it happens is what turns a frown into a smile.
The narrative shift he wants at a societal level is the same one: from replacement to augmentation, making the human better rather than replacing them.
On keeping up, he relies on agency partners and on regular meetings with technology partners, alongside industry conferences and published material.
He is candid that the information environment itself is unreliable, describing two well-placed sources giving him opposite accounts of the same commercial relationship and leaving him to work out where the truth sits.
What the industry has not settled
His first ask of the research is a marketplace of shared practice. He judges that no such thing exists yet for marketing, only ad hoc points of view rather than an aggregate picture of trends and lessons.
He also wants to know where risk sits, and distinguishes between companies making AI an agenda item in itself and those asking what needs solving and whether AI is the tool to solve it.
His second ask is governance. He believes the industry needs principles ensuring transparency, a commitment to truth, and standards for how AI is applied.
He points to a competitor that labels its AI-generated ads, and asks what happens at scale. When a company produces hundreds of thousands of social assets a year, something will slip past human quality checks, so disclosure may be the safer default.
A further unresolved question is ownership. If assets are created by AI, he asks who actually owns them.
His closing point is about trust. Principles are needed so that AI is not used to manipulate people’s thoughts, attitudes or feelings when they cannot tell what is real.
Scope and limitations
The source is an automatically produced meeting transcript with occasional garbled passages, and it records one marketing leader’s view of his own company rather than measured results.
Source
This page summarises Transcript: Brad Audet (Mazda) Interview on Humanized Growth in the Era of AI, by Mariana Peneva, Brad Audet, March 13, 2025.