The whole module on one page — three pillars, the definition, two lists worth memorising — and the AI ad that started the fight of 2024.
July 2026 · Case 1 of 6
As you read — hold this question
When AI can make the next thousand versions of an ad for almost nothing, what actually becomes scarce — and what does “good” marketing mean then?
56M+views
of Coca-Cola’s 2024 AI Christmas ad — and it was reviled and repeated at the same time.
Coca-Cola remade its classic Holidays Are Coming ad entirely with AI in November 2024. It drew tens of millions of views and instant “soulless” backlash — and yet the company did it again in 2025. Cultural failure and commercial success, at once. That tension is the whole module in miniature: the upside of AI marketing is real, and so are its costs.
The standard picture
The textbook versionQuiz: Definitions & benefits
The map: three pillars, one definition, two lists worth memorising
AI in marketing is taught across three pillars — marketing (making the content), sales (working out who will buy), and customer experience (talking to the customer directly). The set definition is deliberately plain: using AI to develop, execute, and improve your marketing (Salesforce framing).
Faster decisions · better return · sharper measurement · stronger relationships · real insight from your data.
The genuinely new thing is not that machines can make content — it is that they can make the next thousand pieces for almost nothing. Generative AI now produces every format: text (copy, emails, scripts), images, video, and voice (narration, dubbing), across a tool landscape the deck lists as roughly twelve products (Canva, Synthesia, Murf and others).
The set example, and what happened next
The rest of the story
The Coca-Cola AI ad — the fight of 2024
In November 2024 Coca-Cola remade its classic Holidays Are Coming Christmas ad entirely with AI. It drew 56M+ views and instant backlash — viewers called it “soulless.” The sharpest line came from Gravity Falls creator Alex Hirsch: “Coca-Cola is ‘red’ because it’s made from the blood of out-of-work artists.”
But the easy “push-button” take is wrong. By the studios’ own account the spot took around twenty people and weeks of work across three studios (Secret Level, Silverside AI, Wild Card) and four generative models. Founder Jason Zada shrugged off the critics — “the spot tested really well” — and the real tell: Coca-Cola did it again in 2025. It failed culturally and worked commercially at the same time.
The other side — the TED talk’s catch
The set TED talk (Jessica Apotheker) leans on a “right-brain / left-brain” creativity metaphor. That is a documented neuromyth — a brain-imaging study (Nielsen et al., 2013, PLOS ONE, n=1,011) found no whole-brain hemisphere bias. The business point survives; the neuroscience does not. The same talk carries two ~40% figures, and usually reports only the happy one: AI can lift creative output ~40%, but over-reliance cuts the divergence of ideas by ~40% — consistency bought as sameness at scale.
Quick recall — without looking back
Test yourself on this case
Question 1 of 3
Define AI-powered marketing, and name the three pillars.
AI-powered marketing is using AI to develop, execute, and improve your marketing (Salesforce framing). The three pillars are marketing (making the content), sales (predicting who will buy and drafting outreach), and customer experience (the machine talking directly to the customer).
Question 2 of 3
List the five impact areas and the five IBM benefits.
Five impact areas: content, personalisation, customer experience, productivity, scale. Five IBM benefits: faster decisions, better return, sharper measurement, stronger relationships, and real insight from data.
Question 3 of 3
The Coca-Cola AI ad ‘failed’ yet Coca-Cola repeated it. What is the exam-ready reading of that, and what is wrong with the TED talk’s brain metaphor?
It failed culturally (backlash, ‘soulless’) but worked commercially (56M+ views, tested well, repeated in 2025) — both true at once. The TED talk’s right/left-brain framing is a neuromyth (Nielsen et al., 2013); and its two ~40% figures cut both ways — +40% output, −40% idea divergence.