A survey of 500 senior marketing and technology executives, which Acquia circulated in London on September 22, 2026 after a first release in Boston on September 14, finds that the two functions give sharply different accounts of who decides on the content platforms companies are now opening to AI systems. Among technology chiefs, 57% described themselves as the decision maker on those platforms, while marketing chiefs spread the same responsibility across three different owners.

In Short

Acquia, a company that sells software for running websites, asked 250 marketing bosses and 250 technology bosses in the United States, the United Kingdom and France who picks the systems that hold a company's web content, and the two groups gave clashing answers. This matters because AI assistants increasingly read those systems to describe brands to shoppers, yet only 41% of the marketing bosses thought their teams could put proper controls on AI-made marketing. The result is an accountability gap that the survey measures but does not settle, because it records what executives believe rather than how their companies actually work. The research comes from a vendor that sells a fix for this exact problem, and the same text went out in Boston on September 14 before the London version followed on September 22.

Three owners on one side, one on the other

The study is titled "The Missing Success Factor for Growth in the Agentic Era: A New Mandate for CMOs and CIOs". Its central finding concerns the software that stores and publishes a company's web content: the content management system (CMS) and the broader digital experience platform (DXP) built around it. Asked who owns decisions on those platforms, chief marketing officers split three ways, according to Acquia. A quarter named marketing, 24% named IT and 18% pointed to a head of digital.

Chief information officers and chief technology officers painted a different picture. According to Acquia, 57% of them said they were the decision maker on CMS and DXP choices and 66% said they were the key evaluator. When the same technology executives described the marketing side, they placed CMOs and marketing directors in the decision-making role 33% of the time and in the evaluator role 42% of the time.

Acquia reads that contrast as evidence that marketing and IT are "describing two different decision-making structures inside the same organisations." The figures support a narrower conclusion.

Start with the arithmetic. The three marketing answers add up to 67%, and the release does not say how the remaining third of CMOs responded - whether they described shared ownership, named another executive or declined to answer. On the technology side, the evaluator shares total 108%, which is possible only if a respondent could assign the role to more than one function. That points to two different response formats: something close to a single choice for marketing, and something closer to a multi-select grid for IT. Acquia describes the two as the same question. The release's own summary list adds to the confusion, stating that "33% of CMOs say they are the decision maker by CIOs' account" - a sentence that attributes one answer to both groups at once. The body of the release makes clear that the 33% and 42% figures came from technology respondents.

The sampling raises a further issue. According to the methodology, the CMO survey and the CIO/CTO survey drew on separate pools of 250 respondents each. Nothing in the release indicates that the marketing and technology executives worked for the same companies. The "same organisations" framing is therefore an inference drawn from two sets of aggregate answers, not a comparison of paired responses inside individual firms. At 250 respondents, the conventional margin of error for a proportion near 50% is roughly six percentage points in either direction, and that calculation assumes a random sample, which a commissioned executive panel is not. The one-point gap between marketing and IT in the CMO answers sits well inside that range.

None of this erases the pattern. Technology executives, taken as a group, see platform decisions concentrated in their own function; marketing executives, as a group, see them dispersed. Whether that reflects disagreement within companies or simply two vantage points on two different sets of companies, the survey cannot say.

A machine as the first reader

Chris Tranquill, chief executive of Acquia, framed the stakes around how brands are encountered. "Trust used to be earned one visitor at a time," he said. "Now it's judged instantly, by a machine, on behalf of someone who may never see your homepage. You don't win that kind of trust with a bigger marketing budget or a stricter compliance checklist. You win it when everything an AI agent finds about your brand comes from one governed source of truth, built by marketing and IT together."

The premise that software now forms a first impression on a buyer's behalf has support beyond Acquia's own data. A Contentsquare survey of 2,000 consumers in the United States and France, released on September 2, 2026, found that 21% had come across AI details contradicting a brand's own page, and 23% would switch brands when the website following an AI recommendation disappointed them. An audit of AI recommendations in August found that one in nine businesses OpenAI recommended carried a wrong claim, while an IAB framework dated August 3 recorded that only 16% of brands systematically track their visibility in AI systems. In June, IAB Australia argued that agents weigh brands on data points rather than creative impressions, so a brand whose information is not structured for machine reading may not be considered.

Such findings turn a procurement question into a question of brand liability. When an AI assistant misstates a price, a product claim or a regulatory disclosure, the error often traces back to content that someone published, or failed to correct, inside a content management system. Who owns that system when something goes wrong?

Where marketing's doubts concentrate

According to Acquia, 66% of CMOs said that demonstrating their AI-generated marketing is governed and controlled has become a brand-reputation issue rather than purely an IT matter. Only 41% believed their own teams could build real governance into their AI initiatives. The company says that was the weakest score among all the priorities it tested for 2027. It does not say how many priorities were tested, and names just one comparison: 73% of CMOs expressed faith in their ability to drive revenue growth.

The explanation Acquia offers is infrastructural. According to the company, 61% of CMOs said their content is spread across so many systems that no single, reliable source of truth exists, and only 40% said their technology stack is ready for AI agents to work inside it at present. Six in ten, in other words, describe stacks not yet ready for AI agents.

Similar gaps appear in other research. A Mediaocean survey of 320 respondents conducted in November 2025 ranked difficulty connecting AI insights across multiple systems at 41% among barriers to AI adoption, just behind data quality or access issues at 42%. StackAdapt's research in August found 49% of marketers citing fragmented data pipelines as a barrier to delegating decisions to AI, with only 19% reporting tools fully integrated into their workflows. ProGEO.ai's index in April, drawn mostly from marketers at cybersecurity companies, found 76.8% of respondents working under a formal generative AI usage policy but only 43.8% at organisations enforcing it with technical controls. Governance on paper is common. Enforcement is not.

IT's conditions for deployment

Technology leaders show the same hesitation from the opposite direction. According to Acquia, 70% of CIOs and CTOs called data sovereignty essential to their organisation, and 56% said they would not deploy AI capabilities until content structure and data integrity are solid. Yet 51% also described migration complexity as a significant barrier to fixing that foundation. Acquia characterises the result as caution and the capacity to act on it pulling in opposite directions.

The release does not define data sovereignty, and the term carries different weight across the three markets surveyed. Amazon Web Services brought its European Sovereign Cloud to general availability in January 2026, with more than €7.8 billion of planned investment in the German deployment by 2040. On July 31, 2026, the chair of the European Data Protection Board asked the European Commission to assess whether a US Supreme Court judgment affects the EU-US Data Privacy Framework.

For the French respondents, fieldwork overlapped with a regulatory milestone. Article 50 of the EU AI Act, which sets transparency obligations for AI-generated content, became legally binding on August 2, 2026, with fines of up to 3% of worldwide annual turnover. Because the release offers no country breakdown, there is no way to tell whether French executives answered differently from their American or British counterparts.

The platforms at the centre of the dispute are also changing what they allow software to do. WordPress.com added write capabilities on March 20, 2026, giving AI agents 19 new tools to draft posts and manage pages through the Model Context Protocol. Cloudflare's EmDash, a content management system covered in April, includes a built-in MCP server in every instance. Once agentic AI can edit a live site rather than merely read it, platform ownership becomes a question of who signs off on changes no human typed.

The return on investment question

The anxiety running beneath both functions is financial. According to Acquia, 43% of CMOs named failure to prove a return on AI investment as the single biggest threat to their position as a marketing leader in 2027, ahead of missing revenue targets or losing budget to IT. The release gives no percentages for those alternatives.

Other data suggests the fear is grounded. TransUnion research published on August 5, 2026, drawn from 100 senior US marketing leaders, found only 53% reporting meaningful return on AI investment while 89% expected to increase AI spending, with 65% measuring AI's impact mainly through estimated time or cost savings. KPMG's Global AI Pulse, published on June 24, 2026 from a survey of 2,145 senior leaders, found that 49% had narrowed, delayed or paused AI agent deployments when costs outran value, and only 7% had reached established return on investment. Gartner forecast in June 2025 that more than 40% of agentic AI projects would be cancelled by the end of 2027, citing escalating costs, unclear business value and inadequate risk controls.

Budget structure adds pressure. Gartner's 2026 CMO Spend Survey, fielded among 401 marketing leaders between January and March 2026, found that martech's share of marketing budgets had fallen to 19.4% from 26.6% in 2021, while AI initiatives took 15.3%. A CMS or DXP contract sits squarely inside that shrinking martech line. If IT increasingly makes the platform call, as the technology executives in Acquia's sample claim, the prospect of losing budget to IT is less hypothetical than the release's ranking implies.

What the methodology leaves out

According to Acquia, the findings come from independent research commissioned by the company, conducted by Coleman Parkes Research and analysed by VCCP. Fieldwork ran between July and August 2026. Each survey - 250 CIOs and CTOs in one, 250 CMOs in the other - covered the United States, the United Kingdom and France.

Several details are absent. The release gives no breakdown by country, company size, industry or revenue band. It publishes no question wording. It does not say whether CIOs and CTOs, two distinct roles, were pooled or reported separately. The full report sits behind a registration form on Acquia's website; the landing page describes sections on four types of CMO approach to agentic marketing and five proposed next steps for marketing and technology leaders.

The chronology also differs from the London dateline. Acquia's resource centre lists the report with a date of September 9, 2026. The company's US release, carrying the same headline, findings and quote, was distributed through GlobeNewswire from Boston at 08:00 Eastern Time on September 14, 2026. The London edition dated September 22 repeats that text with British spelling.

The commercial context is plain. Acquia was co-founded in 2007 by Dries Buytaert, who began the Drupal project as a community message board at the University of Antwerp in 2000, according to the company. Acquia describes itself as the world's number one Drupal hosting provider and sells Acquia Source, which it calls a digital command centre where teams manage content, applications, AI agents and analytics in one place. Its website also markets separate offerings for governing AI-generated content and for improving brand visibility in AI answer engines. A study concluding that marketing and IT need one governed source of truth, built together on a single platform, points directly at the category of product its sponsor sells. That does not invalidate the data, but it explains the framing.

One customer, one site

Acquia offers Edrington, the spirits group behind The Macallan and Highland Park, as evidence that governance and growth need not be traded off when marketing and IT build on the same platform. According to the release, Edrington worked with Acquia partner FreelyGive to take "a similar AI-native approach" to a stalled website migration, using AI to generate and test website components directly within Acquia Source, including a compliance-critical age gate, without slowing down its editors. The word "similar" has no antecedent in either version of the release; no other example precedes it.

Acquia's customer video page for the project supplies more detail, and a narrower scope. According to that page, the property in question was a B2B site within a portfolio of smaller brand and campaign sites, managed by a team in the Dominican Republic and previously running on WordPress with GoDaddy hosting. Acquia says the project reached user acceptance testing within weeks at a fraction of the earlier cost, without giving a figure, and that the team rebuilt the age gate using AI and prompt engineering during a single 15-minute call.

What the case does not show is the subject of the survey. There is no indication of which executive owned the platform decision at Edrington, no traffic or compliance outcome and no cost figure. It illustrates build speed on a single site. It does not demonstrate a resolved accountability structure.

Why the ownership gap reaches the marketing plan

The London version of the release lands as British industry data points to more formal structures than Acquia's CMOs describe. IAB UK research published in May 2026 found that 57% of businesses in digital advertising had internal governance or ethics structures covering AI use, rising to 76% among larger firms, although that report did not detail what those frameworks require. DigiCert's survey of 1,001 IT and cybersecurity decision makers, published on July 7, 2026, found that 78% of organisations had experienced an AI-related security incident or vulnerability while roughly half lacked formal governance. Structures exist in many companies. Whether they reach the content systems that AI assistants read is a separate matter.

For marketers, the Acquia numbers describe a practical problem. The content management system feeds landing pages, product claims, structured data and, increasingly, the answers AI assistants assemble about a brand. Agentcy's AI Visibility Index, based on 104 senior B2B marketing leaders surveyed in February 2026, found that 81% considered AI visibility a blind spot in their organisation.

The survey's most defensible finding may be its simplest. Marketing leaders doubt they can control what AI produces in their name, and technology leaders say they will not switch it on until the underlying content is in order. Both conditions point at the same asset. Which executive answers for that asset when an AI system gets a brand wrong is the question the research raises, and the one its own data, drawn from two separate pools of respondents, is not built to answer.

Timeline

Summary

Who: Acquia, the Boston-based Drupal hosting and digital experience platform company, and its chief executive Chris Tranquill, reporting on 250 CMOs and 250 CIOs and CTOs surveyed by Coleman Parkes Research, with analysis by VCCP.

What: A commissioned survey finding that technology executives largely see themselves as the decision makers on CMS and DXP platforms, with 57% claiming the role, while CMOs split ownership between marketing (25%), IT (24%) and a head of digital (18%). Only 41% of CMOs believed their teams could build real governance into AI initiatives, 61% said their content lacks a single reliable source of truth, 56% of technology executives said they would not deploy AI until content structure and data integrity are solid, and 43% of CMOs named unproven AI returns as the biggest threat to their position in 2027.

When: Fieldwork ran between July and August 2026. The report is listed on Acquia's site with a September 9, 2026 date, the US release went out on September 14, 2026, and the London edition is dated September 22, 2026.

Where: Respondents were based in the United States, the United Kingdom and France. The release was issued from Boston and later from London.

Why: Content management systems increasingly supply the information AI assistants and agents use to describe brands, which makes platform ownership a question of accountability for AI errors. The survey's methodology, including separate respondent pools, apparent differences in question format and undisclosed country splits, limits how far its central claim of disagreement inside the same organisations can be taken.