The AI Capability Gap

THE BRIEF

Seventy percent of CMOs say becoming a leader in artificial intelligence is a primary goal for 2026. Only 30% believe their organizations have the infrastructure to do it. That 40-point spread — documented in Gartner's annual CMO Spend Survey, fielded across 401 senior marketing leaders at companies with revenues exceeding $1 billion — is the defining number of this moment in marketing leadership. The aspiration is nearly universal. The execution capacity is not. For CMOs in the next budget review, this is the conversation that determines whether they're leading the transformation or absorbing the blame when it underdelivers.

Marketing budgets haven't moved. According to the same Gartner data, average marketing spend held at 7.8% of company revenue in 2026, up from 7.7% in 2025 — a flat line dressed as stability. The catch: organizations with optimized AI programs are averaging 8.9% of revenue in marketing budgets. There is a measurable premium already attached to AI execution capability. CMOs who can scale AI aren't just winning operationally — they're capturing budget share their peers are leaving on the table. The gap between the prepared and the unprepared is beginning to show up in resource allocation, not just results.

The talent problem compounds the infrastructure problem. A separate study from the World Federation of Advertisers and Ogilvy Consulting — drawing on 80 senior marketers across 57 companies representing more than $60 billion in combined ad spend — found that while 78% of companies have made AI and technology a primary focus of their marketing transformation, only 12% of marketing teams possess the necessary expertise to successfully leverage AI. In other words: the mandate is set, the budget is being directed, and nearly nine in ten marketing organizations are running the play without the players capable of running it.

Consumer trust in AI-generated advertising is eroding faster than most marketing leaders have accounted for. Research from Canva (disclosed: vendor-produced) found that 70% of consumers say they can "tell for the most part" when an ad is AI generated — and that it is "missing its soul." More materially: 74% report being more likely to purchase from an ad made entirely by humans versus one generated by AI. This is not a sentiment signal. It is a conversion signal. CMOs scaling AI content production without a corresponding quality and authenticity layer are likely paying an invisible tax on their conversion rates right now without seeing it reflected in their attribution models.

At the same time, there's a proof point that AI — deployed correctly — changes the economics of marketing fundamentally. Hershey's VP of Brand Development and Brand Creation Strategy, Daniel Mohnshine, reported this week that the company has reduced its innovation pipeline development time by three months — at the concepting stage — by using AI to accelerate consumer feedback and prototyping cycles. For a company operating global supply chains and retailer relationships, a 90-day compression in a product's path from concept to test is not an efficiency win. It restructures the competitive math. Brands that move faster through the concepting phase can run more experiments, fail earlier, and bring only validated ideas forward to production.

OpenAI launched a self-serve ChatGPT Ads Manager this week, opening its ad platform to U.S. businesses with tools for budget setting, campaign management, cost-per-click bidding and conversion measurement. The platform is simultaneously expanding to the U.K., Mexico, Japan, Brazil and South Korea. Target, Albertsons and Williams-Sonoma have been early pilot participants. OpenAI's agency integration already includes Dentsu, Omnicom, Publicis and WPP. The company projects $2.5 billion in ad revenue this year and has publicly stated a target of $100 billion by 2030. ChatGPT is no longer a platform where marketing leaders get to wait and watch. It is a paid media channel. The planning question is no longer "should we?" It is "how much, and under what governance?"

---

---

---

THE SIGNAL

The tension: Every boardroom wants AI transformation in marketing. Most marketing organizations are structurally unequipped to deliver it. The CMOs navigating this are not failing from lack of ambition — they are failing from lack of infrastructure while being measured as if the infrastructure were already in place.

This is not a skills gap story. It is a performance accountability story with a skills gap underneath it.

The Gartner CMO Spend Survey, released this week, makes the split stark: 70% of CMOs call AI a key 2026 priority; only 30% feel their organizations have the infrastructure to scale it. The gap is 40 points — wide enough to be structural, not situational. Simultaneously, 56% of CMOs report lacking budget to execute their 2026 strategy and 54% report lacking the necessary resources. These are not the same CMOs hedging — they are marketers being set a target they have not been given the tools to hit.

The organizations winning are pulling away precisely because they solved for infrastructure before they scaled for ambition. Gartner's data reveals that marketing organizations with optimized AI programs allocate an average of 21.3% of their marketing budgets to AI — versus 15.3% for the average. They are also receiving a higher revenue share: 8.9% versus 7.8%. The compounding advantage here is material. Better-equipped teams direct more budget to AI, which produces better results, which justifies more budget. The CMO who is still building the foundation is watching this gap widen in real time.

The WFA and Ogilvy data adds precision to where the gap lives. Ninety-six percent of multinational companies report being in active transformation. Seventy-eight percent have made AI and technology a primary driver. But only 12% of marketing teams have the expertise to execute. That is not a rounding error — it is a structural workforce deficit that cannot be closed by buying more software or hiring one AI lead. It requires systematic investment in upskilling that most marketing organizations are not making. Gartner's February data reinforces the problem: while nearly two-thirds of marketers believe AI will fundamentally change their jobs, only 32% believe they need to update their skills to respond. The majority are expecting transformation without preparing for it.

Who is winning and why. The organizations pulling ahead share a specific pattern: they have separated AI-as-workflow-tool from AI-as-strategic-capability. Hershey's is the clearest current example — using AI earlier in the innovation cycle, in concepting and consumer feedback, not just in content production. The result is not faster ads. It is a compressed development timeline that changes how many bets they can run per quarter. That is a competitive advantage with a direct margin implication: more validated ideas reach production, and fewer expensive commitments are made on unvalidated concepts.

The organizations falling behind are deploying AI primarily at the content layer — generating more assets faster — without addressing the data foundations that would allow those assets to be properly targeted, measured and attributed. Canva's research (vendor-disclosed) shows 85% of marketers reporting AI is saving them four or more hours per week. But if that time saving is happening in content creation while the data and measurement infrastructure remains immature, the productivity gain is real and the strategic gain is not.

Who is losing. The CMOs losing ground in this cycle are those whose AI investment is primarily defensive — adopting tools because the board asked them to, without a coherent thesis for how AI connects to revenue growth or margin improvement. The Drum's coverage this week captures the pattern precisely: "performative fluency" is widespread — marketing leaders nodding in vendor meetings, presenting AI strategies they cannot fully execute, building confidence that erodes quietly as delivery fails to match declaration. According to Fiona McKenzie, Managing Director of Marketbridge, this gap creates a structural organizational problem: slower decision-making, over-reliance on specialists, and misalignment across marketing, product, data and commercial teams — each holding different interpretations of what their AI investments are actually enabling.

The ChatGPT inflection. OpenAI's launch of a self-serve Ads Manager this week creates a new competitive divide on a different axis. ChatGPT, with its reported 800 million weekly active users as of early 2026, is now a managed paid media channel with cost-per-click bidding, conversion measurement and an expanding global footprint. The brands that will extract the most from this channel in the next 12 months are those that already understand how to reach audiences in conversational contexts — a fundamentally different behavioral environment than display or search. CMOs who treat ChatGPT advertising as an extension of their existing search or programmatic playbook will underperform against those who approach it as a new intent architecture entirely.

The stakes. The timeline here is not abstract. Gartner has stated that half of agencies' proprietary AI platforms will be obsolete by 2029. That is three years. CMOs who spend 2026 and 2027 in performative AI adoption — without building the underlying infrastructure — will face the double pressure of obsolete tooling and no institutional capability to adopt what replaces it. The organizations investing now in data foundations, AI governance, and systematic skills development are not just winning in 2026. They are building the compounding advantage that determines market position in 2029.

---

---

THE DEEP DIVE

Thesis: The AI capability gap in marketing is a structural problem, not a talent problem — and closing it requires a different set of investments than most CMOs are currently making.

The surface reading of this week's data is familiar: CMOs want AI, organizations aren't ready. The more useful reading is why they aren't ready — and the answer is not "people aren't skilled enough." It is that most marketing organizations were built for a fundamentally different operational model, and they are trying to run AI on infrastructure designed for something else.

The infrastructure mismatch. Gartner's chief of research in its Marketing practice, Ewan McIntyre, framed it precisely: "CMOs recognize AI's potential as a force multiplier for growth, efficiency and transformation, but most marketing organizations are not yet built to capture that value. The risk is that CMOs invest in AI tools faster than they build the data foundations, processes, governance and talent required to scale them." The sequence matters. AI tools at scale require: clean, unified data; governance frameworks that define who can do what with AI and under what review; integration between AI outputs and existing tech stacks; and measurement systems capable of attributing AI-driven results separately from other channels. Most organizations are buying the tools before any of those conditions are met.

The Salesforce State of Marketing (10th edition, disclosed: vendor-produced, ~4,500 respondents) identifies the data problem specifically: 83% of marketers recognize the shift toward personalized, two-way engagement — but only one in four are satisfied with how they use data to power those interactions. That 75% dissatisfaction rate on data utility is not a marketing problem. It is a data architecture problem that marketing is being asked to solve with tools that require the architecture to already exist.

The practitioner reality. The gap between what AI can theoretically deliver and what it delivers in practice shows up clearly in how practitioners talk about it. The emerging practitioner vocabulary is instructive: "AI slop" — AI-generated content perceived as generic or lazy — has become a recognized problem inside marketing organizations. In Canva's research (vendor-disclosed), 41% of marketing leaders describe AI slop as "a considerable challenge." This is a pipeline quality problem, not a creative problem. When AI is used to scale content production without a corresponding quality review layer, volume increases and brand coherence decreases. The consumer signal confirms this: 70% of consumers report they can detect AI-generated advertising, and the response in comment sections — as noted by Citi's Jonathan Harrop, VP of Marketing and Communications for Wholesale Lending, this week — is "massively negative." That sentiment is not contained to social media. It affects purchase behavior: 74% of consumers report being more likely to purchase from fully human-made ads (Canva, 2026, vendor-disclosed).

The skills paradox. The data on skills investment reveals a specific form of institutional denial. Nearly two-thirds of marketers believe AI will change their jobs — but only 32% believe they need to update their skills to respond (Gartner, February 2026). This is the gap between intellectual acknowledgment and behavioral response. Teams are processing that transformation is coming without changing how they work today. The WFA and Ogilvy Consulting data draws the line: 78% of organizations have AI and technology as a primary transformation focus. 12% of marketing teams have the expertise to execute on it. The organizations where AI investment is actually paying off — like Hershey's, which compressed its innovation pipeline by three months — are not necessarily hiring AI specialists. They are building teams that use AI to amplify skills they already have, earlier in processes where it creates compound value.

The framework for closing the gap. CMOs looking at this data should resist the temptation to see this as a tools problem with a software solution. The pattern in organizations that are executing effectively points to four specific investments:

1. Data foundation first. AI tools are only as useful as the data they process. Organizations with 21.3% AI budget allocation (Gartner) are not spending more on tools — they are spending more on data quality, integration and governance that makes tools work. Before the next AI platform purchase, the diagnostic question is: does the data feeding it accurately represent actual customer behavior at the resolution required to make decisions?

2. Process integration before content scaling. Hershey's three-month compression is not the result of generating more AI content. It is the result of integrating AI earlier in the innovation process — at the concepting and consumer feedback stage — where it changes the decision logic, not just the production cost. CMOs should identify where in their existing processes AI could change what decisions get made, not just how fast.

3. Quality layer as non-negotiable. The consumer data on AI content detection suggests that organizations scaling AI content production without a human review standard are paying an invisible conversion penalty. Brand guidelines, voice standards and creative quality criteria need to apply to AI outputs with the same rigor they apply to agency deliverables.

4. Skills as a structural budget line, not a training event. The 12% expertise rate (WFA/Ogilvy) will not close through periodic workshops. It closes through deliberate role design — building AI fluency expectations into job descriptions, reviews and promotion criteria — and through sustained investment in upskilling that is tracked and measured like any other operational metric.

The failure mode to watch. The most dangerous failure mode in AI marketing adoption is the credibility decay that results from visible AI errors — brand consistency breaks, content that consumers identify as generic, personalization that misfires — in organizations where the quality layer was never built. These failures do not arrive dramatically. They arrive through incremental brand trust erosion that only shows up in longer-cycle metrics: net promoter scores, repeat purchase rates, brand preference measures. By the time those numbers move, the damage has been accumulating for months. The CMOs who are watching those downstream metrics for early AI failure signals are ahead of those waiting for performance dashboard alerts.

The consequence. The organizations that close the capability gap in the next 18 months — by building data foundations, integrating AI into strategic processes (not just content pipelines), maintaining quality standards and investing systematically in skills — will enter 2028 with a compounding structural advantage. They will have more validated institutional knowledge about what AI-driven marketing actually produces, at what cost, under what conditions. That knowledge is not replicable by a competitor who waits until the technology matures. The advantage is in the institutional learning, not the tooling. The CMOs who treat 2026 as the year to build that foundation — not just to declare an AI priority — are making the decision with the longest payoff horizon.

---

---

THE PLAYBOOK

C-SUITE

• Audit the gap between AI declaration and AI execution by asking your CMO for three specific proof points where AI investment has changed a business outcome (pipeline, conversion, product cycle) — not efficiency metrics alone — because the Gartner data shows the CMOs producing revenue results are allocating 40% more budget to AI infrastructure, not just tools, and the board question should be whether your organization is investing in the right layer.

• Pressure-test your marketing org's data readiness before the next AI platform approval, because Salesforce's data shows 75% of marketers are dissatisfied with how they use existing customer data to power personalization — and AI tools layered on top of bad data infrastructure produce bad AI outputs at scale.

• Set a 2027 benchmark for AI-driven revenue attribution, not AI adoption rates, because organizations that have optimized AI programs are already receiving 1.1 points more revenue share in marketing budgets and the compounding advantage will be measurable at the P&L level within 18 months.

CMO / VP MARKETING

• Separate AI-as-content-production from AI-as-strategic-infrastructure in your budget allocation and report on them differently to the board, because the organizations winning are allocating 21.3% of marketing budgets to AI (vs. the 15.3% average) specifically toward data foundations and process integration — not toward tools that generate content faster without solving the upstream measurement problems.

• Build an explicit quality review standard for all AI-generated customer-facing content before the next content scaling initiative, because 70% of consumers now report they can detect AI-generated advertising and 74% are more likely to purchase from human-made ads — meaning brand content that reads as AI-generated is carrying a conversion penalty your current attribution models probably aren't isolating.

• Identify one high-cycle process — innovation concepting, campaign briefs, audience segmentation — where AI can be integrated to change what decisions get made (not just how fast), because Hershey's compressed a product development cycle by three months by applying AI earlier in its innovation pipeline, and the competitive advantage of that compression is not in the speed but in the number of validated bets the business can now run per year.

DEPARTMENT LEADS

• Require AI literacy as a performance criterion in your next team review cycle and define what "AI fluency" actually means for each role, because the WFA and Ogilvy data shows only 12% of marketing teams currently have the expertise to execute AI strategy — and the skills deficit will not close through voluntary training without structural accountability attached to it.

• Establish a clear escalation path for AI quality failures (content that breaks brand standards, personalization errors, AI-detectable creative) before those failures accumulate into brand trust damage, because the practitioner signal from multiple sources this week is that "AI slop" is a recognized and growing problem inside marketing organizations, and teams without a defined quality standard are generating it without a reliable mechanism to catch it.

• Create a 90-day proof-of-concept for ChatGPT advertising with explicit measurement criteria before the channel becomes a default budget line, because OpenAI's Ads Manager is now self-serve and major agency partners — Dentsu, Omnicom, Publicis, WPP — are already integrated, meaning the channel will move from pilot to standard faster than most planning cycles anticipate.

---

---

THE NUMBERS

Source: Gartner CMO Spend Survey, January–March 2026 (401 CMOs, $1B+ revenue organizations, North America, Europe, U.K.)

Source: Gartner CMO Spend Survey, 2026

Source: Gartner CMO Spend Survey, 2026

Source: Gartner CMO Spend Survey, 2026

Source: Gartner CMO Spend Survey, 2026

Source: WFA and Ogilvy Consulting, 2026 (80 senior marketers, 57 companies, $60B+ combined ad spend)

Source: Canva, 2026 (DISCLOSED: Vendor-produced research)

Source: Daniel Mohnshine, VP Brand Development and Brand Creation Strategy, The Hershey Company, via The Drum, May 11, 2026

Source: Axios, April 2026; confirmed in Marketing Dive reporting, May 11, 2026

—

THE DARK CALLOUT:

70% of CMOs have made AI a 2026 priority. 30% have the infrastructure to execute it. The board doesn't know that split exists. Yet.

—

---

WHAT'S NEXT + WHAT'S COMING

The signal gaining the most momentum across practitioner channels this week is the widening gap between AI adoption metrics and AI outcome metrics inside marketing organizations — surfacing simultaneously in Gartner survey data, trade press interviews and practitioner commentary. The CMO conversation is shifting from "how do we adopt AI?" to "how do we prove it's actually working?" — and most organizations don't yet have the measurement infrastructure to answer that question. Watch for attribution and ROI measurement to become the defining capability differentiator in the back half of 2026; the CMOs who can connect AI spend to revenue outcomes with specificity will have leverage in every budget conversation their peers will not.

One specific thing to watch before next Tuesday: OpenAI's Ads Manager is expanding to the U.K., Mexico, Japan, Brazil and South Korea. Watch for the first wave of case study data from early pilot participants — particularly from Target and Albertsons — to determine whether ChatGPT advertising is performing as a discovery channel, a conversion channel, or primarily a brand awareness vehicle. That distinction will determine how the channel is ultimately budgeted and measured.

M&A, TOOL LAUNCHES & EVENTS:

• OpenAI launched ChatGPT Ads Manager (self-serve) — U.S. beta, expanding to five additional countries. CPC bidding and Conversions API now live. (May 11, 2026)
• Heineken announced it is consolidating its agency roster around "fewer future-fit partners" — a signal that major brands are tightening agency relationships as AI capabilities shift the value equation in client-agency work. (Marketing Dive, May 2026)
• Gartner predicts 50% of agencies' proprietary AI platforms will be obsolete by 2029 — watch for agency consolidation and platform pivots accelerating in H2 2026.
• Trade Desk cited AI chatbot advertising as a key future opportunity following a difficult Q1, signaling the independent buy-side is positioning for the ChatGPT/AI channel wave before it arrives at scale.

---

## SOURCES

1. Gartner CMO Spend Survey 2026 (401 CMOs, January–March 2026, $1B+ revenue orgs, North America/Europe/U.K.) via Marketing Dive, May 11, 2026: https://www.marketingdive.com/news/ai-remains-a-top-priority-for-cmos-but-spending-lags-gartner/819806/

2. Gartner, "CMOs want AI transformation, but few are upgrading their skills," February 2026, via Marketing Dive: https://www.marketingdive.com/news/gartner-cmos-want-ai-transformation-but-few-are-upgrading-their-skills/812450/

3. World Federation of Advertisers (WFA) and Ogilvy Consulting, 2026 research (80 senior marketers, 57 companies, $60B+ combined ad spend), via Marketing Week, May 11, 2026: https://www.marketingweek.com/transformation-time-saving-consumer-confidence-5-interesting-stats/

4. Canva, 2026 research (DISCLOSED: vendor-produced), via Marketing Week, May 11, 2026: https://www.marketingweek.com/transformation-time-saving-consumer-confidence-5-interesting-stats/

5. Salesforce, "State of Marketing," 10th Edition (DISCLOSED: vendor-produced, ~4,500 respondents worldwide), 2026: https://www.salesforce.com/marketing/resources/state-of-marketing-report/

6. Daniel Mohnshine, VP Brand Development and Brand Creation Strategy, The Hershey Company, via The Drum, May 11, 2026: https://www.thedrum.com/news/hershey-s-daniel-mohnshine-on-how-ai-is-cutting-product-development-by-three-months

7. Jonathan Harrop, VP Marketing and Communications, Citi, via The Drum, May 11, 2026: https://www.thedrum.com/news/inside-the-jury-room-citi-s-jonathan-harrop-on-why-ai-is-dragging-cmos-back-into-the-weeds

8. Fiona McKenzie, MD, Marketbridge, via The Drum, May 11, 2026: https://www.thedrum.com/opinion/we-ve-made-marketing-leaders-accountable-for-technology-we-never-taught-them-to-understand

9. OpenAI ChatGPT Ads Manager announcement, May 7–11, 2026, via Marketing Dive, May 11, 2026: https://www.marketingdive.com/news/openai-solidifies-ad-platform-ambitions-with-chatgpt-ads-manager/819801/

10. OpenAI ad revenue projections via Axios, April 2026: https://www.axios.com/2026/04/09/openai-100-billion-in-ad-revenue

11. Marketing Week, "Most CMOs say they lack budget to deliver their 2026 strategy," May 11, 2026 (citing Gartner): https://www.marketingweek.com/cmos-lack-budget-strategy/

---

This report was produced with AI assistance and human editorial review.