AI For Growth, Not Just Efficiency: A CMO Playbook
Many brands still treat AI like a cost-cutter. Useful, yes-but it leaves growth, creativity and strategic leverage on the table. The AI Advisory Council for CMOs (formed with Revmatics) views this moment as a shift point: pair technical literacy with bold experimentation to rethink how brands communicate, serve and grow.
As Zena Arnold, CMO at Sephora, put it: "We've talked a lot about how we can use AI for cost savings. I'd love to think about using it to deliver growth and to deliver value to our end customers." Mike Benson, president and CMO at CBS, reinforced the mindset: "AI isn't here to take your job. It's here to help you do your job better."
The Council brings together leaders from Ally Financial, AT&T, Chime, Coca-Cola, e.l.f. Beauty, Instacart, JPMorganChase, Mars Petcare, Mastercard, PepsiCo Foods, Samsung, Stellantis and more-focused on frameworks that connect experimentation to business outcomes.
Demystifying Data: Build Once, Scale Everywhere
The "golden dataset" is your multiplier
Vineet Mehra, CMO at Chime, was clear: "Data is the fuel that makes AI truly work the way we all want it to." Translation: your brand guidelines, voice, content and customer signals must live in one trusted system. Without this foundation, models drift, outputs vary and risk increases.
Ekta Chopra, chief digital officer at e.l.f. Beauty, warned against fragmented pilots: "If you don't think about a unified data strategy upfront, that's the biggest challenge." Chime built a shared "golden dataset" used across legal, compliance, marketing and product-so every model trains on the same truth.
- Structured data: CRM, transactions, site analytics. Without clean, labeled, accessible fields, AI can't act.
- Unstructured data: "Video, text, audio is our richest, most underutilized resource," said Ricky Ray Butler, co-founder and CEO of Revmatics. With contextualized datasets, models deliver creative and performance gains.
- Synthetic data: Artificially generated "digital twins" to test creative, targeting and attribution-without exposing real consumer data. "Synthetic data is so underutilized," said Raja Rajamannar, chief marketing and communications officer at Mastercard.
The advantage goes to brands that can cut across all three types, build accurate profiles and decide which messages and offers fit the moment.
Operating Models That Actually Scale
Three out of four executives expect gen AI to materially change marketing operating models within two years. Source: McKinsey's 2024 AI survey (read more).
Standardize inside, be partner-agnostic outside
Romain Mallard, global VP of marketing, network operations and capabilities at the Coca-Cola Company, summed up the approach: define how you work internally, then plug in the best partners for each link in the value chain. That includes agencies, freelancers, tech companies and APIs-whoever does the job best without disrupting the whole.
New roles, new guardrails
Nearly three-quarters of marketing orgs expect their talent mix to change due to gen AI. Emerging roles include AI content strategists, data curators and ethical AI officers. Mastercard even appointed a chief risk officer for marketing and communications.
As Raja Rajamannar noted: products commoditize fast. The differentiator becomes innovation and creativity-enabled by AI and executed through new ways of working.
Bridge Innovation With Practical Use Cases
1) Customer value: solve real problems
Former Progressive Insurance CMO Remi Kent challenged leaders to "stay focused on AI's human purpose." That means designing systems that augment people and reflect values-so experiences get more relevant, trusted and helpful.
Jennifer Brockington, executive director of marketing strategy, media and innovation at Ally Financial, added: "Marketers' true value is to take AI outputs and translate that into actionable brand and product innovations." Marketers become the interpreters-turning model insights into useful features and moments that move both emotions and numbers.
AT&T chief marketing and growth officer Kellyn Smith Kenny emphasized anticipation: serve today's needs and predict tomorrow's. Leading teams are reworking workflows so marketing, product and CX co-develop offerings at the speed of culture-using AI to sense behavior shifts, simulate outcomes and adapt before trends peak.
Accuracy and safety matter. "AI is changing how consumers search," said Rajoielle Register, senior VP and chief marketing performance and growth officer at Stellantis. "You can't just unleash the beast and risk spreading misinformation." Chime's internal, secure LLM-trained on real member insights-lets teams query like they're talking to a customer and move faster on creative and product decisions.
2) Community management: consistency at scale
e.l.f. trained a private LLM in AWS on brand guidelines and social history, with humans approving each post. The result: response rates jumped, and 90% of comments now get answered in brand-approved language. That's scale with control.
3) CRM efficiencies: compress cycle time
Instacart compares outputs from multiple internal models to match each task with the best engine. They cut CRM production timelines by 80%. "Maintain control of the data environment while encouraging responsible experiments," said CMO Laura Jones. The pattern: decentralized pilots, centralized guardrails.
Agentic AI: From Automation To Orchestration
Agentic systems can reason, collaborate and coordinate actions between steps-shifting AI from "generate" to "decide and do." Early adopters report strong returns; Bain & Co. details the shift to agentic operating models (overview).
Council members previewed systems coordinating audience intelligence, creative development and media activation in real time. Tracy-Ann Lim, global chief media officer at JPMorgan Chase: "Agentic AI is shifting marketing from automation to orchestration." The goal is to bring teams and partners together around real-time response to audience behavior.
Public tools (Amazon's Rufus, ChatGPT agents) are moving agentic patterns into shopping. Early signals look promising: traffic lifts and better experiences for those who try agentic shopping. Still, there are limits.
Chase Zieman, co-founder and CTO at Revmatics, called out a practical issue: the more you ask a single model to do at once, the lower the quality. So you split tasks into agents, then you need to design, orchestrate and monitor the system. Security company Human reports only 2.2% of agents touch cart, checkout and payment pages-meaning humans still close the loop. Bain's take: most companies aren't ready; systems, data and governance must evolve first.
Build The Foundation Now
Chris Bellinger, chief creative officer at PepsiCo Foods U.S., framed it well: we're just starting to see the creative and operational upside of AI. The next wave belongs to leaders who invest in data, talent, use-case alignment and clear guardrails.
A 6-step operating checklist
- 1) Define value: Choose 3-5 use cases that affect revenue, margin or retention. Rank by impact x feasibility.
- 2) Build your golden dataset: Centralize structured, unstructured and synthetic data with governance, consent and PII policies.
- 3) Standardize the stack: Document workflows, access controls and evaluation methods. Be partner-agnostic on tools.
- 4) Pilot with guardrails: Human-in-the-loop review, bias checks, brand voice QA and secure environments.
- 5) Measure what matters: Time-to-market, cost-to-serve, conversion lift, LTV, CAC payback, brand trust signals.
- 6) Upskill the org: Train marketers to prompt, interpret and productize model outputs. Add roles: data curator, AI content strategist, ethical AI officer.
90-day starter plan
- Days 1-30: Stand up a cross-functional AI council. Select use cases. Inventory data. Define policies. Choose pilot tools.
- Days 31-60: Build a minimal golden dataset. Ship two pilots (e.g., community responses, CRM content). Set evaluation baselines.
- Days 61-90: Expand to a third use case tied to revenue. Integrate human review. Document wins and gaps. Plan scale-up.
This is the first thought leadership piece from the AI Advisory Council for CMOs. A detailed white paper, created with Ad Age Studio 30 and Revmatics, will cover how brands can move from automation to adaptation with real-world frameworks and use-case models.
Helpful resources
- AI Certification for Marketing Specialists - practical skills, brand-safe workflows, measurable outcomes.
- AI courses by job - curated learning paths for marketing, product and analytics teams.
Your membership also unlocks:
