Alex signed up for a 10K trail race and asked an AI assistant for a training plan and shoe recommendations suitable for flat feet and a history of knee pain. In seconds, the assistant produced a tailored training schedule, compared three shoe options, and summarized hundreds of reviews — while filtering out irrelevant mass-marketing offers and flagging a retailer with poor return policies. It even completed the purchase with next-day delivery and a convenient return option. This kind of AI-enabled shopping experience is already happening.
This article is authored by Eli Stein (McKinsey Bay Area), Jamie Wilkie (Boston), Julien Boudet (Southern California), Kelsey Robinson (Boston), and Lalit Bhagia (Mumbai), with Emily Scofield representing McKinsey’s Growth, Marketing & Sales Practice. It was edited by Barr Seitz (New York).
How consumer behavior is shifting
McKinsey surveys show nearly half of consumers already use AI-based search to guide purchase decisions (McKinsey AI Discovery Survey, August 2025, n = 1,927). Shoppers today use twice as many channels on average to inform or make purchases than they did ten years ago, and as agents become more common, shopping will grow faster and more agentic.
This shift is changing marketing at a structural level: the campaign-era model is becoming inadequate.
AI capabilities and marketing impact
AI systems analyze data, detect patterns, and make predictions or decisions. Generative AI creates new content (text, images, video, code). Agentic AI combines AI with generative models to plan, decide, and execute across workflows with limited human input. These capabilities affect marketing in distinct ways, from faster content production to continuous, self-optimizing campaigns.
Where organizations stand today
Marketers are optimistic — 86 percent say they are excited about AI — and 90 percent of CMOs are experimenting with AI use cases. Yet less than 10 percent have scaled AI or captured value across marketing workflows. A major reason is that many organizations apply AI as point solutions; only 28 percent are pursuing a fundamental rewiring of teams and workflows.
When done well, McKinsey client experience suggests AI can deliver 4–7 percent revenue growth, two- to threefold productivity improvements, and 60–70 percent savings on execution-related tasks. McKinsey estimates the global B2C retail market could generate $3–5 trillion in value from AI.
Five capabilities of an AI-first marketing engine
- Continuous insights
- Translate signals from customers, markets, and channels into real-time decisions.
- Generative AI reduces costs through automation, fewer errors, and faster timelines.
- Example: digital twins that simulate consumer personas to test campaign, pricing, and product responses.
New role: Customer Wayfinder — synthesizes data and insights, tests ideas with synthetic audiences, provides cultural understanding, and applies strategic judgment.
- Scaled creativity
- Produce large volumes of tailored content while maintaining brand consistency.
- Agents can monitor LLMs, search, and social ecosystems to detect trends and intent shifts, then generate, test, and optimize content for humans and AI agents.
- Many organizations start their AI journey here; senior marketers rate generative content production as their most mature AI capability.
Impact: some organizations see 2–5x creative productivity and 10–30 percent reductions in creative costs; campaign cycles can compress from weeks to same-day execution.
New role: Creative Guru — defines systems and guardrails for content creation, ensures brand consistency, and uses performance data to improve outputs.
- Hyperpersonalization
- Continuously learning systems deliver one-to-one experiences in real time across channels.
- AI-driven personalization can increase customer satisfaction by 15–20 percent, revenue by 5–8 percent, and reduce cost-to-serve by up to 30 percent.
New role: Hyperpersonalization Architect — designs data models, AI capabilities, and business rules for one-to-one experiences and oversees compliance and trustworthiness.
- Marketing to AI agents (agentic commerce)
- Influence how AI systems interpret and recommend products or services.
- Agentic experiences are conversational, adaptive, and continuously learning; the economy is shifting from attention to trust as recommendation systems increasingly determine brand choice.
- More than half of consumers already rely on AI for purchase guidance, putting significant portions of traditional web search traffic at risk.
Companies must be not only visible but also “consumable” by machines: machine-readable knowledge, credibility signals (detailed specs, verified reviews, expert input), and up-to-date information are essential.
New role: Agent Whisperer — ensures accurate brand representation in AI systems and that agentic programs build value and consumer trust.
- Always-on orchestration
- Replace campaign cycles with continuously managed and optimized marketing run by human–agent teams.
- Properly configured, always-on orchestration can improve marketing ROI by about 30 percent and cut time spent on execution tasks from 60–70 percent to 10–15 percent.
New role: Full-funnel Navigator — oversees the entire marketing system, sets agendas, guides agentic systems, shapes algorithmic inputs, and uses AI-driven insights to optimize performance.
Client examples and measurable outcomes
- A leading consumer-technology company rebuilt marketing into an AI-first growth engine with reusable marketing agents: ~35–50 percent time savings in campaign activation, roughly 20 percent reduction in external spend, and compression of AI-driven content and audience generation processes from 10–12 weeks to minutes.
- A financial-services company invested in structured data and a machine-readable knowledge engine and rewired high-value workflows so AI agents could generate hypotheses, test formats and narratives, and dynamically reallocate budgets; humans decided which signals to prioritize. Within nine months the company increased organic traffic sixfold while improving customer quality, lowering acquisition costs, and materially increasing retention.
How to get started
Key priorities for executives:
- Rewire high-value marketing workflows: break work into tasks and decide which parts people or AI agents should perform; establish technical foundations (unified data/identity layers), end-to-end KPIs, and cross-functional teams.
- Build a new human–AI hybrid organization: redesign roles, incentives, and training. New roles will appear for builders, orchestrators, and quality stewards; skills development is the top barrier cited by senior marketers.
- Invest in tech and data foundations at scale: agentic platform architecture, modular interoperable frameworks, robust data pipelines to feed enriched, structured data into LLMs, and revised partnership models for third-party AI platforms.
- Ensure governance and active value capture: implement rigorous testing standards, clear decision rights, and a dynamic transformation office that sets explicit value goals from day one and rapidly reallocates resources when needed.
AI automation can save time, but savings only create value if freed capacity is deliberately redeployed into growth activities.
Conclusion
The future of marketing will be defined not by the number of AI tools organizations deploy but by how well they connect insight, creativity, personalization, agentic commerce, and execution into a single, continuously improving system. For CMOs, the immediate task is to design that system so AI’s potential converts into measurable business outcomes.
Authors and acknowledgments: Eli Stein (partner, Bay Area), Jamie Wilkie (partner, Boston), Kelsey Robinson (senior partner, Boston), Emily Scofield (associate partner, Boston), Julien Boudet (senior partner, Southern California), and Lalit Bhagia (partner, Mumbai). The authors thank Fahmi Hamzah, Joe LaRose, Shihui Mao, and Varun Mehrotra for their contributions. The article was edited by Barr Seitz, editorial director, New York.
