Why Humans Succeed in Optimizing Accessibility

Automated tools can catch common accessibility code violations, but what about the human experience? To date, they cannot replicate the authentic, complex experience of manual testing done by humans. This is especially true for those with diverse physical or cognitive challenges.

Yet, the speed of these tools is undeniable. And that’s why pairing the strengths of automated tools with human testing and feedback looks to be the best means of creating a truly inclusive product.

As people and businesses become increasingly reliant on digital properties, failure to provide access is not an option. Lost customers equate to lost revenue.

Let’s get right to the techniques and testing methods necessary to deliver fully accessible products for diverse user groups, using devices with or without assistive additions, and in multiple potential environments. 

Collaborative accessibility testing — humans and machines unite

Most accessibility issues evade automated testing tools, but a hybrid approach infusing human manual testing helps produce an accessible product for all user types. Test automation or machines can provide initial accessibility scans — and they can be effective at it.

The benefits of using automated tests for initial accessibility testing include: 

• finding 20-40% of general accessibility issues, according to several accessibility services providers, including Applause in its 2025 State of Digital Quality in Accessibility
• covering approximately 25% of WCAG Level A criteria and 17% of WCAG Level AA criteria, according to accessibility testing studies cited by Applause and the Royal National Institute of Blind People
• accelerating testing when combined with human accessibility testing and review, per the W3C Web Accessibility Initiative

But there are limitations, as humans are far more likely to find accessibility issues like: 

• errors in context or the meaning of text 
• interpreting standards for changes and nuanced meanings
• poor experiences for those using assistive devices
• inconsistencies or inefficiencies in the accessibility process itself, which result in downstream issues

Humans excel in understanding human customers. Machines excel at scanning code for specific types of back-end errors or coding mistakes.

Accessibility testing technique musts

Human manual testing must cover a variety of areas, such as content review, keyboard shortcuts and screen reader accuracy. It’s imperative that organizations also test assistive device functionality — and with so many devices in market today, this is a difficult challenge. At a minimum, human accessibility testing should validate the following areas: 

• Zooming in up to 200%
• Keyboard entry works
• Correct UI placement and readability
• Functioning buttons, forms, and menus
• Body text
• Unique page titles
• Descriptive text is included 
• Tags like <h1> are present — and only once on the page 
• Logical headings maintain programmatic <h1> through <h6> order
• Acronyms are spelled out on first reference 
• Text excludes jargon or undefined acronyms
• Images
• Every image includes accurate descriptive text
• Confirm text is not embedded into images
• Color
• Text and background image or color has contrast ratios at a minimum of 4.5:1
• Confirm color is not used to convey meaning on its own
• Forms 
• Form controls include descriptive text labels
• Tables
• Tables must not be used as layout tools 
• When data tables are used, verify table captions, rows and columns all have headers with text.

Perform keyboard verification tests, including the following actions: 

• Selecting the Tab key to navigate buttons, links and form controls in a logical order
• Pressing Shift + Tab should navigate backward 
• Using the spacebar or Enter buttons to activate links and buttons
• Navigating with the arrow keys to select radio buttons, drop-down menus, sliders, tab panels, auto-complete fields and menus 
• Pushing the Esc. key should close browser dialog windows and menus

Testing screen readers requires understanding how they work. Screen readers convert text from mobile or computer screens into a usable format for users who cannot read it. Screen readers typically convert text to speech or sound. Users adjust the voice speed, pitch and verbosity. For blind users, screen readers convert text to Braille along with optional audio. 

Screen reader users navigate websites using keyboard commands more often than by mouse clicks. Special shortcut keys enable them to navigate a page faster than by listening to audio of every element. 

When testing screen readers, the reading order is important. Screen readers move through a page based on the order of elements in the underlying code rather than display order. If the code doesn’t match the visual sequence, text becomes confusing or difficult to follow. Web and mobile pages must follow a logical reading order and move through elements and content linearly.

To support the more than 1.3 billion people living with disabilities worldwide — and to stay compliant with evolving regulations like WCAG 2.1 and the EAA — engineering teams need testing workflows that move as fast as their release schedules. Here is where Applause and its testing services can support internal teams on that journey.

Applause offers a fully managed strategy that combines expert assessments with real-world feedback from people with disabilities — and it pays off. By shifting accessibility testing earlier in the SDLC, organizations can pair rigorous accessibility reviews guided by professionals with in-sprint testing executed by community testers using authentic assistive technologies — like screen readers, switch controls and magnification tools. This real-device validation delivers clear, prioritized bug reports and code-level remediation guidance directly into the backlog, enabling engineering teams to fix critical usability barriers quickly. This is how many innovative brands turn accessibility from a post-production patch into a proactive part of a release cadence.

Adversarial areas affecting accessibility

The AI era introduced an unexpected differentiator for many brands — sites that are accessible to humans are also easily AI readable. Images with meaningful descriptions are useful for blind users and multimodal AI models attempting to learn how to associate language with images. Content structured for human accessibility is also better positioned for AI accessibility. 

People with disabilities are often the unintended victims of conflicts between digital publishers and AI tool developers. As business, technology and accessibility change, our ethical responsibility to provide a full digital life to all users means building inclusive digital worlds. After all, accessible websites and apps benefit all end users. When organizations deliver inclusive, accessible products, we all win.