AI-assisted marketing needs stronger accessibility controls

Marketing teams can create copy, landing pages, personalization and code faster with AI-assisted tools. That increases the amount of customer experience they can put into production. Accessibility defects can scale with that output when controls fail to catch them. Marketing therefore needs quality controls that keep pace with production.

AI speed changes what marketing is accountable for

Accessibility often enters workflows after marketing has defined an experience. Engineering or compliance may then handle technical accessibility requirements. But marketing also influences the CMS, AI platforms and production tools that determine what reaches customers. Decisions made before an engineering handoff can shape whether accessibility problems repeat across campaigns and pages.

The available figures do not establish that AI caused a deterioration in web accessibility. The more useful management issue is production scale. When teams generate or modify more digital experiences in less time, an unresolved problem can spread through more pages, campaigns and customer journeys before detection. Executives need controls designed for the volume and frequency of current production.

Some failures are difficult to detect through visual review alone. A page can look correct while its structure remains unclear to a screen-reader user, or while someone navigating by keyboard cannot complete a form. Visual approval answers only part of the quality question. Marketing also needs evidence that customers can navigate the experience and complete its intended tasks.

Accessibility was already widespread; greater scale raises the stakes

The 2026 WebAIM Million report found detectable accessibility failures on 95.9% of the top one million website homepages, with an average of 56.1 errors per page. It reported that detected errors increased 10.1% from the prior year, reversing six consecutive years of gradual improvement. For executives, these figures make accessibility a widespread production-quality issue.

WebAIM also reported that the average homepage contained 1,437 elements, a 22.5% increase in a single year and nearly twice the level in 2019. More elements create more places where labels, controls, interactions and content relationships may require accessibility checks. The simultaneous rise in complexity and detected errors gives executives reason to examine quality controls, while establishing no causal link between AI and either change.

Accessibility problems also predate the current generation of AI production tools. WebAIM identified low-contrast text, missing alternative text, unlabeled form fields, empty links, empty buttons and missing document language among the six most common failure types across seven consecutive years. Faster production can reproduce these persistent defects when teams fail to catch them before release.

A June 2025 New York City Bar Association report stated: “AI cannot solve for accessibility if it was never trained to recognize it.” The statement raises a procurement and testing question rather than establishing the capabilities of every AI system. Teams need to test what a particular system produces and which accessibility failures its controls can detect.

Automation provides partial accessibility coverage

Automated accessibility tools can repeatedly scan pages and flag known machine-detectable problems. This matters when a company publishes frequently or manages many digital properties. Detection and automated remediation provide evidence for a defined set of checks. Broader assurance requires evaluation of issues that depend on context, interaction and human judgment.

Reported estimates put the limitation in concrete terms: automated tools can detect about two-thirds of accessibility issues and fix roughly half of the issues they detect. Meaningful uncertainty remains after a scan, even at those levels. Executives should define what each tool tests and establish how teams evaluate the remaining issues.

Consider a campaign form that appears correct on screen. Automated testing may identify a missing programmatic label or another known code-level defect. Other accessibility questions depend on context, interaction and whether a customer can complete the full task. Human testing and expert review can address those questions in workflows that matter to customers.

Litigation also tests the assumption that installing a tool resolves accessibility exposure. AudioEye’s 2026 Web Accessibility Litigation Report found that 38.5% of companies facing accessibility lawsuits already had an accessibility tool when they were sued. AudioEye sells accessibility products and services and benefits commercially from demand for stronger accessibility programs, so executives should assess its findings with that incentive in view.

AudioEye also reported several measures of U.S. digital-accessibility litigation:

A count of combined federal and state claims is a different measure from a count of lawsuits. Keeping those terms distinct matters when executives assess changes in legal exposure.

Reported settlement costs typically range from $15,000 to $75,000, with legal fees of $10,000 to $30,000 on top. These costs can make an accessibility failure a direct financial issue alongside remediation and staff time. Legal exposure is one consequence executives can include when governing accessibility risk.

The legal issue also crosses national markets. On June 4, 2026, a French court ordered Carrefour to bring carrefour.fr and its app to full accessibility conformance within six months. The ruling was described as the first under the European Accessibility Act to reject partial conformance as a legal defense, while Carrefour’s own accessibility statements had claimed 50% to 70% conformance across its platforms.

For management, the key question is what evidence shows that an experience works for customers. Automated checks can provide part of that evidence and reduce manual workload, while expert evaluation and user testing address issues requiring context or human judgment. Governance determines which checks happen before release and how teams handle unresolved findings.

Partial coverage becomes a customer problem

Accessibility gaps affect customers directly. A screen reader converts digital information into speech or other output, allowing a person to navigate and understand an interface without relying on its visual presentation. Keyboard navigation lets someone move through links, controls and forms without a mouse. A page that looks finished to a campaign team can still prevent a customer from finding information or completing a purchase.

Reported customer-behavior figures say seven in 10 customers with disabilities leave websites they find inaccessible, while 83% restrict their shopping to sites they already know work for them. If verified, those behaviors connect accessibility directly to marketing performance. Barriers can reduce the audience that reaches conversion.

More than 1.3 billion people globally live with disabilities and, together with their family and friends, represent over $18 trillion in spending power, according to the Return on Disability Group. These figures frame accessibility as a customer and market issue alongside legal exposure. Inaccessible experiences can limit who is able or willing to respond to marketing investment.

Marketing teams can measure accessibility work through existing performance processes. A team can establish baseline conversion and bounce rates on a high-traffic page, address identified accessibility problems, then measure the same outcomes afterward. Accessibility issue counts can sit alongside customer acquisition cost and conversion rate during performance reviews. Changes in those metrics still require careful interpretation because several factors can affect customer behavior at the same time.

Accessibility belongs inside marketing’s production system

The ownership change starts before content or code reaches production. When marketing participates in selecting a CMS, personalization engine, code generator or other AI-powered platform, accessibility should be part of the procurement criteria. Buyers can ask which problems the system detects, which it can remediate, whether human review is involved and who handles cases beyond its automated coverage. Speed and accessibility performance each need evidence.

The same principle applies to vendors that combine automation with services. AudioEye describes its approach as combining AI-powered automation with expert-written custom fixes. Because AudioEye sells accessibility solutions, this is the company’s characterization of its own offering and supports its commercial interests. Buyers can apply the broader principle across vendors: identify the boundary of automation and assign responsibility for work beyond it.

During production, automated tests can provide fast feedback while a page is still easy to change. Teams can scan templates, campaign pages and recurring components as they are created. Fixing a machine-detectable failure before publication can also prevent the defect from carrying into later work. Repeated checks matter when AI-assisted systems increase the frequency of changes.

Manual and expert testing can concentrate on areas where software provides weaker assurance. High-value paths such as account registration, search, forms and purchasing deserve evaluation as complete experiences, including tests with screen readers and keyboard navigation. The goal is to determine whether customers can perform the intended task. A clean visual review or completed automated scan provides only part of the evidence required for that judgment.

Accessibility can then become a launch criterion. A landing page, product launch or seasonal campaign can have defined accessibility checks before release, an owner for unresolved findings and a clear rule for issues that block publication. Marketing influences the launch decision and can prevent known barriers from reaching customers. This makes accessibility a production control earlier in the workflow.

The control continues after launch because digital experiences change. Content edits, personalization, new components and generated variations can alter an experience that previously passed review. Continued monitoring can catch detectable regressions, while periodic human evaluation can test whether important customer journeys still work as intended. Continuous production requires accessibility checks throughout the operating cycle.

Engineering remains essential to implementation. Technical teams may own component architecture, semantic code, remediation and platform-level fixes that marketing cannot perform itself. Marketing controls or influences procurement, production, optimization and release, giving it responsibility for accessibility decisions at those stages. Shared ownership places each control with the team that can act on the relevant decision.

Key executive takeaways

• Match accessibility controls to AI production speed: Marketing’s faster content and experience production can reproduce accessibility defects at greater scale. Build accessibility checks into procurement, production and release workflows.
• Treat accessibility as a production-quality issue: WebAIM found detectable failures on 95.9% of the top one million homepages in 2026 as page complexity also increased. Marketing and technology owners need controls that catch recurring defects before they spread.
• Define the limits of automation: Automated tools cover only part of the accessibility problem, leaving issues that require context, expert review and user testing. Map each tool’s coverage and assign responsibility for unresolved findings.
• Connect accessibility to customer performance: Inaccessible forms, navigation and purchasing journeys can prevent customers from completing the actions marketing intends. Track accessibility findings alongside conversion, bounce and other performance measures while accounting for other factors that affect results.
• Build accessibility into marketing operations: Marketing owns or influences technology selection, campaign production, optimization and launch decisions. Set accessibility criteria for procurement, test critical journeys before release and continue monitoring experiences as they change.