Build an Agile AI Marketing Organization

1. The AI Growth Gap: Dave’s CMO research exposes a strategic mismatch. Most chief marketing officers expect AI and generative AI to drive growth, and 60% of surveyed companies planned to invest more than $10 million during the following year. Yet their principal use cases remained manual-work reduction, content creation, employee productivity, faster production, and asset consistency. Those activities matter, but they do not automatically alter how customers perceive or use a brand. Closing the gap requires leaders to treat AI as a mechanism for redesigning customer journeys and value propositions, not simply as a faster production layer. Productivity should become the foundation for growth investment rather than the final ambition.

2. Three Levels of AI Value: Dave separates AI value into three levels. “For my productivity” covers individual uses such as drafting emails or job descriptions with ChatGPT and creating content with Jasper. “For our efficiency” applies AI to end-to-end functional performance, including faster campaign delivery and finer measurement. “For how we compete” uses AI to change customer experiences and value propositions. The distinction matters because each level requires different sponsorship and evidence. Individual adoption can be informal; process efficiency requires coordinated workflows; competitive differentiation demands senior leadership and cross-functional action. At Aetna, for example, more than 100,000 contracts were in motion, illustrating the scale at which AI can create operational value before customer-facing transformation begins.

3. Measure From a Real Baseline: Platform claims such as “saved a day” are weak unless buyers understand how the estimate was calculated. Dave recommends establishing a baseline before implementation. With a pharmaceutical company pursuing agile, personalized marketing, the team examined its project management system and measured how long the previous five campaigns took to reach market. New work was then compared with that record and proved dramatically faster. A split test offers stronger evidence by comparing AI-enabled and conventional approaches simultaneously. A before-and-after comparison is less statistically clean, but it remains useful when controlled testing is impractical. The principle is direct: organizations cannot credibly show lift if they have not documented where performance started.

4. From Campaigns to Programs: Traditional campaigns are batch-oriented pushes: a launch, discount, or quarterly promotion is created and then targeted to reduce wasted spend, perhaps with two or three segment variants. Dave contrasts this with an ongoing program that asks what the next-best touch should be for each customer at a given moment. Programs monitor customer data and prospect activity for triggers, respond according to journey context, and continuously test and improve. Search engine marketing already operates this way through ongoing optimization. Extending that model across marketing changes planning, staffing, measurement, and governance. The objective is no longer merely to get an asset out the door; it is to operate a persistent learning engine.

5. AI Changes Team Design: Marketing specialization emerged partly because platforms such as social networks, search engines, websites, and programmatic buying systems developed distinct operating mechanics. One of Dave’s clients required 13 people to get a marketing program out the door. AI increasingly automates those platform idiosyncrasies, reducing the premium on narrow execution expertise. Dave sees more value moving toward strategic questions: what role should social or search play, at which stage of the journey, and how should resources shift as results emerge? The resulting model uses multidisciplinary teams of five or six people who can perform more than one activity and work in rapid test-and-learn cycles. AI supports the model, but broader human capability makes it viable.

6. Insourcing Will Reshape Agencies: AI will make routine marketing production easier to bring in-house. Dave cites an Association of National Advertisers survey showing that more than 70% of members had established some form of internal agency capacity. More than 80% of that group expected to expand it substantially over the next three years. This does not eliminate external agencies: companies will still seek major ideas, specialist expertise, and additional capacity. However, tier-two and tier-three work such as routine content adaptation and day-to-day execution will become faster and cheaper internally. Agencies therefore need sharper differentiation, while corporate marketing leaders need operating models that combine internal speed with selective external creativity and expertise.

7. Reskilling Must Be Structural: Interwell Health illustrates how organizations can prepare people for broader AI-enabled roles. Its marketing team has introduced tools including Jasper, Microsoft Copilot, and a video platform while expecting employees to become more versatile. The organization allocates five hours per month to AI learning and discusses AI familiarity, interest, and perspective with job candidates. The requirement is not daily expertise with every tool; it is willingness to learn and work differently. Dave emphasizes that people are not born knowing how to use AI. Training must therefore be built into schedules, plans, and leadership expectations. Otherwise, role expansion feels like added workload rather than an opportunity to increase impact.

8. Reallocate People Toward Experience: An insurance company working with Dave expanded its CMO role into a chief experience officer remit spanning marketing and the post-purchase journey. The company needed more people to manage customer experience initiatives, including onboarding, but did not have materially more budget. Its solution was to cap the content production team, use AI to increase content output, and direct incremental headcount toward customer experience management. This is a useful workforce principle: automate scalable production where tools are capable, then invest scarce human capacity in ambiguous, strategic work. AI may eventually execute parts of an onboarding design, but people must first determine what the experience should be and coordinate the organization to deliver it.

9. Buy-In Requires Visible Proof: At Aetna, Dave asked the head of Medicare marketing to replace a predefined campaign plan with an agile agreement: set the growth target and budget, then allow the team to test and learn toward the outcome. Her accounting background made the uncertainty uncomfortable, so she approved an initial two-week trial and could place a compliance representative alongside the team. Dave then brought her into the Medicare marketing “war room,” where cross-functional staff and agencies reviewed results, moved money, and selected new tests. Seeing the process changed her view so decisively that she asked for it to be filmed for peers. The lesson is that transformation gains support through controlled exposure, evidence, and governance.