Ability to Access
Ability to Access
Ability to Access
Rethinking application approach of WCAG 2.2 accessibility guidelines.
The Web Content Accessibility Guidelines (WCAG 2.2) are the global standard for accessible digital products. In practice, they are applied one success criterion at a time, to individual components. As a result, a screen can meet every criterion and still fail the people using it. Ability to Access reorganises the WCAG 2.2 success criteria around the human capabilities they protect and the types of tasks they apply to. The framework was then built into a Claude AI skill that audits screens and live websites across six task types, reports design and development audit.
Organisation:
frog
Duration:
12 weeks
Team:
Sanskruti Bansal, POOJa, punnya kr, pavani rastogi
Capabilities:
Design research, Accessibility design, UX auditing, Framework design, Human-AI interaction
Rethinking application approach of WCAG 2.2 accessibility guidelines.
The Web Content Accessibility Guidelines (WCAG 2.2) are the global standard for accessible digital products. In practice, they are applied one success criterion at a time, to individual components. As a result, a screen can meet every criterion and still fail the people using it. Ability to Access reorganises the WCAG 2.2 success criteria around the human capabilities they protect and the types of tasks they apply to. The framework was then built into a Claude AI skill that audits screens and live websites across six task types, reports design and development audit.
Organisation:
frog
Duration:
12 weeks
Team:
Sanskruti Bansal, POOJa, punnya kr, pavani rastogi
Capabilities:
Design research, Accessibility design, UX auditing, Framework design, Human-AI interaction
Rethinking application approach of WCAG 2.2 accessibility guidelines.
The Web Content Accessibility Guidelines (WCAG 2.2) are the global standard for accessible digital products. In practice, they are applied one success criterion at a time, to individual components. As a result, a screen can meet every criterion and still fail the people using it. Ability to Access reorganises the WCAG 2.2 success criteria around the human capabilities they protect and the types of tasks they apply to. The framework was then built into a Claude AI skill that audits screens and live websites across six task types, reports design and development audit.
Organisation:
frog
Duration:
12 weeks
Team:
Sanskruti Bansal, POOJa, punnya kr, pavani rastogi
Capabilities:
Design research, Accessibility design, UX auditing, Framework design, Human-AI interaction
Back
The catalyst
We audited Aadhaar, world’s largest biometric ID system at the National School for the Blind in Bengaluru.
Despite formally meeting WCAG AA standards, real users faced intense cognitive load. For example, They could not locate key services. forced into a long trial-and-error route due to dense menus, complex terminology, and unlabelled controls.
The catalyst
We audited Aadhaar, world’s largest biometric ID system at the National School for the Blind in Bengaluru.
Despite formally meeting WCAG AA standards, real users faced intense cognitive load. For example, They could not locate key services. forced into a long trial-and-error route due to dense menus, complex terminology, and unlabelled controls.
The catalyst
We audited Aadhaar, world’s largest biometric ID system at the National School for the Blind in Bengaluru.
Despite formally meeting WCAG AA standards, real users faced intense cognitive load. For example, They could not locate key services. forced into a long trial-and-error route due to dense menus, complex terminology, and unlabelled controls.
Insight
Guidelines are applied one at a time, to isolated widgets
On the Aadhaar OTP screen, one guideline passes while two others fail on the same task. Checked separately, nobody sees that the person still cannot log in.
Insight
Guidelines are applied one at a time, to isolated widgets
On the Aadhaar OTP screen, one guideline passes while two others fail on the same task. Checked separately, nobody sees that the person still cannot log in.
Insight
Guidelines are applied one at a time, to isolated widgets
On the Aadhaar OTP screen, one guideline passes while two others fail on the same task. Checked separately, nobody sees that the person still cannot log in.
Break through
Apply guidelines together, capability wise, across the whole task.
CAPABILITY OVERLAP : One task leans on several abilities at once, so the guidelines that protect them are applied together, not one by on OVERALL TASK EXPERIENCE : Accessibility is judged across the full task flow, from start to finish, not screen by screen or component by component.
Break through
Apply guidelines together, capability wise, across the whole task.
CAPABILITY OVERLAP : One task leans on several abilities at once, so the guidelines that protect them are applied together, not one by on OVERALL TASK EXPERIENCE : Accessibility is judged across the full task flow, from start to finish, not screen by screen or component by component.
Break through
Apply guidelines together, capability wise, across the whole task.
CAPABILITY OVERLAP : One task leans on several abilities at once, so the guidelines that protect them are applied together, not one by on OVERALL TASK EXPERIENCE : Accessibility is judged across the full task flow, from start to finish, not screen by screen or component by component.
Integrating the breakthrough into existing framework. Six capabilities, six task types
To integrate this lens into the existing WCAG accessibility framework, we begin with defining the six core human capabilities required for six fundamental digital task by priority.
Integrating the breakthrough into existing framework. Six capabilities, six task types
To integrate this lens into the existing WCAG accessibility framework, we begin with defining the six core human capabilities required for six fundamental digital task by priority.
Integrating the breakthrough into existing framework. Six capabilities, six task types
To integrate this lens into the existing WCAG accessibility framework, we begin with defining the six core human capabilities required for six fundamental digital task by priority.
84 WCAG 2.2 guidelines, regrouped by ability and task
Next, All the 84 WCAG 2.2 guidelines are regrouped capability wise. This approach allows to design/audit end- to end task with lens of capabilities needed to fullfill them.
84 WCAG 2.2 guidelines, regrouped by ability and task
Next, All the 84 WCAG 2.2 guidelines are regrouped capability wise. This approach allows to design/audit end- to end task with lens of capabilities needed to fullfill them.
84 WCAG 2.2 guidelines, regrouped by ability and task
Next, All the 84 WCAG 2.2 guidelines are regrouped capability wise. This approach allows to design/audit end- to end task with lens of capabilities needed to fullfill them.
AI tool
Translated into an intelligent Claude AI Skill to streamline audit.
Instead of improvised polybags, officers pull standardised pouches from a portable dispenser and cut them to fit any exhibit. The dispenser carries pouches in three sizes.
AI tool
Translated into an intelligent Claude AI Skill to streamline audit.
Instead of improvised polybags, officers pull standardised pouches from a portable dispenser and cut them to fit any exhibit. The dispenser carries pouches in three sizes.
AI tool
Translated into an intelligent Claude AI Skill to streamline audit.
Instead of improvised polybags, officers pull standardised pouches from a portable dispenser and cut them to fit any exhibit. The dispenser carries pouches in three sizes.
Accessibility compliance at a glance
Two charts summarise the whole audit before the detail. Together they show where the page fails, who needs to fix it, and how far each task is from compliance.
Accessibility compliance at a glance
Two charts summarise the whole audit before the detail. Together they show where the page fails, who needs to fix it, and how far each task is from compliance.
Accessibility compliance at a glance
Two charts summarise the whole audit before the detail. Together they show where the page fails, who needs to fix it, and how far each task is from compliance.
Findings are reported task-wise, then grouped by the ability they protect
The report lists findings under each of the six task types. Inside each task, findings are grouped by the human capability they affect, such as Comprehension, Spatial understanding, etc. A designer sees why a task is a success or a fail.
Findings are reported task-wise, then grouped by the ability they protect
The report lists findings under each of the six task types. Inside each task, findings are grouped by the human capability they affect, such as Comprehension, Spatial understanding, etc. A designer sees why a task is a success or a fail.
Findings are reported task-wise, then grouped by the ability they protect
The report lists findings under each of the six task types. Inside each task, findings are grouped by the human capability they affect, such as Comprehension, Spatial understanding, etc. A designer sees why a task is a success or a fail.
A ticket like anatomy of a finding
Each finding is written like a ticket, so it can be verified, assigned and fixed.
A ticket like anatomy of a finding
Each finding is written like a ticket, so it can be verified, assigned and fixed.
A ticket like anatomy of a finding
Each finding is written like a ticket, so it can be verified, assigned and fixed.
Segregation of design and dev audit
The report separates design findings from developer findings, so each team gets its own list and no issue appears twice. If an issue needs both, it sits under the team that must act first and is flagged for the other.
Segregation of design and dev audit
The report separates design findings from developer findings, so each team gets its own list and no issue appears twice. If an issue needs both, it sits under the team that must act first and is flagged for the other.
Segregation of design and dev audit
The report separates design findings from developer findings, so each team gets its own list and no issue appears twice. If an issue needs both, it sits under the team that must act first and is flagged for the other.
The audit states what it cannot test
Some checks need assistive technology, real devices or real users, which an AI working in a browser can’t fully replicate. The report lists every guideline it audited in full, in part, or not at all, and why. That’s what makes it fair and reliable: it never claims a pass it couldn’t verify.
The audit states what it cannot test
Some checks need assistive technology, real devices or real users, which an AI working in a browser can’t fully replicate. The report lists every guideline it audited in full, in part, or not at all, and why. That’s what makes it fair and reliable: it never claims a pass it couldn’t verify.
The audit states what it cannot test
Some checks need assistive technology, real devices or real users, which an AI working in a browser can’t fully replicate. The report lists every guideline it audited in full, in part, or not at all, and why. That’s what makes it fair and reliable: it never claims a pass it couldn’t verify.
The report doubles as a guide to build for accessibility
A designer reading the report knows which task type has issues, which capability it fails, and what to strengthen. Every finding explains what to solve, why it matters to the people it affects, and how to fix it. That equips designers and developers to build for accessibility.
The report doubles as a guide to build for accessibility
A designer reading the report knows which task type has issues, which capability it fails, and what to strengthen. Every finding explains what to solve, why it matters to the people it affects, and how to fix it. That equips designers and developers to build for accessibility.
The report doubles as a guide to build for accessibility
A designer reading the report knows which task type has issues, which capability it fails, and what to strengthen. Every finding explains what to solve, why it matters to the people it affects, and how to fix it. That equips designers and developers to build for accessibility.
