For nearly three decades, I carried ideas about disability employment, accessibility auditing and inclusive technology that I could describe but could not afford to build. Then AI changed the equation.
For much of my life, I have had the feeling that I arrive at ideas before the tools needed to build them arrive.
I could see what was needed.
I could understand the problem.
I could design processes and create practical tools.
But I was not a software developer. I did not know how to turn everything I imagined into code, and I certainly did not have the money to employ development teams every time I saw a problem that technology might solve.
Then AI arrived.
AI did not give me the ideas. It gave me the means to act on ideas I had sometimes carried for decades.
The story began long before AI
My work did not begin with artificial intelligence.
It began with exclusion.
After becoming disabled, I discovered how difficult it could be to participate in a society that simply had not designed itself with people like me in mind. By 1998, when I was documenting the origins of Closing the Gap, I was already describing how ignorance, inaccessible systems and social attitudes could determine whether a disabled person was able to participate at all.
That experience led me toward disability awareness, training and, eventually, employment inclusion.
One idea became central to my work:
Disabled people were not failing systems; systems were failing disabled people.
Employment became one of the first places where I tried to change that.
1998: imagining a digital employment system
I established Closing the Gap because I could see that employment inclusion could not be solved by simply matching a disabled person with a vacancy.
People needed skills.
Employers needed knowledge.
Workplaces needed to be accessible.
Reasonable accommodation needed to be understood.
And barriers had to be removed before a job offer could become genuine participation.
By December 1998, I had prepared a website brief for Closing the Gap that included online forms for job seekers to submit CVs, employers to seek workers, employer subscriptions and job-seeker registration.
I already had an Access database and was asking how it could be used "on the net."
Today, we might call that a two-sided digital employment platform.
I did not.
I simply knew that employers and disabled job seekers needed a better way to find one another.
The wider Closing the Gap model included employment placement, computer training, website development, access consulting and other services. Employers would be able to seek advice about reasonable accommodation and disability employment equity. Job seekers could access employment, skills-development and bursary information. Self-employed consultants could seek work opportunities.
I did not know the phrase digital ecosystem.
But the thinking was already there.
Technology as an equaliser
In 1998, I wrote that computers could help level the playing field for disabled people and open new possibilities in communication, education and employment.
That belief has stayed with me.
Closing the Gap linked computer training to an employment agency so trainees could move from learning toward work. The model combined training, internships, job placement and the identification and removal of barriers.
Technology was never the objective. Participation was.
Technology was simply one of the tools that could make participation possible.
2001: online education before online education became ordinary
In 2001, technology changed what was possible for me personally.
I was accepted to study for an MBA through the University of Liverpool and K.I.T. eLearning. The programme was entirely Internet-based.
At the time, this was highly unusual. A contemporary article described the programme as the first MBA training conducted entirely online, with its first group including students from many countries.
Today, online postgraduate education is normal.
Back then, people were still debating whether serious academic education could really happen through a computer.
I was already doing it.
I received financial support from K.I.T. eLearning and the Department of Labour to make that opportunity possible.
Years later, I would recognise the parallel:
The Internet gave me access to learning. AI would later give me access to building.
2001: turning accessibility standards into practical tools
At around the same time, I was trying to solve another problem.
Accessibility standards existed.
But standards sitting in books did not automatically change buildings.
People needed practical tools.
Closing the Gap developed a structured accessibility checklist using South African accessibility requirements, including SABS 0246 and SABS 0400 Part S, together with international guidance including the ADA.
The system contained separate survey forms for parking, loading areas, routes, ramps, stairs, entrances, doors, lifts, rooms, toilets, signage, alarms, ATMs, restaurants and other elements.
Each of those existed as an individual template.
Today I might describe that as a modular assessment architecture.
Then, I simply thought of it as the practical way the work needed to be organised.
The checklist was intended to help people assess facilities against standards and identify barriers. Importantly, assessors were not simply expected to tick "yes" or "no." They were instructed to record the actual problem and measurements where necessary so that somebody reviewing the evidence later could decide what modifications were required.
That was evidence-based accessibility assessment.
On paper.
Long before I began calling what I do today accessibility intelligence.
I tried to build a business around it
This was not merely theory.
Closing the Gap carried out actual accessibility audits.
One Disability Access Audit for SCAT assessed a building against South African accessibility standards and the ADA. It identified barriers including inaccessible parking, excessive door-opening force, lift-level issues, inaccessible gates, missing signage and entrance barriers.
The toilet area assessed required substantial remodelling.
And the report did not merely identify failure.
It proposed change.
Accessibility assessment should not end with "you failed." It should help answer: "What needs to change?"
The three-legged pot
Over time, I came to describe Closing the Gap through three connected components:
employment equity consulting and training, job placement, and Universal Design.
I had my own name for the model.
I called it the African three-legged pot.
Without one of its legs, I said, the pot would be rather useless.
Years later, AI introduced me to the language of ecosystems.
And suddenly I recognised something about my own earlier work.
I had already built an ecosystem. I simply called it a three-legged pot.
Employment could not work without accessibility.
Accessibility alone did not create employment.
Training did not solve inaccessible workplaces.
An employer could genuinely want to recruit disabled people and still have technology, buildings, recruitment practices or attitudes that excluded them.
Everything affected everything else.
That was the ecosystem.
I had the idea long before I had the language.
My MBA research confirmed what experience had taught me
My MBA dissertation investigated barriers to gainful employment for disabled people in South Africa.
The research examined barriers experienced by disabled job seekers and employees as well as the difficulties employers and recruitment agencies experienced in placing disabled people.
One of the conclusions was critical:
Progressive legislation alone was not enough.
Reasonable accommodation and changes to the surrounding environment were essential.
Again, the issue was systemic.
2015: trying to turn the accessibility model into an app
By 2015, smartphones had changed what seemed possible.
Through Newcastle University's App Movement programme, I initiated an app called Disability accessible facilities.
I wanted consumers to be able to review and rate the accessibility of facilities based on their experience.
I needed 100 people to support the idea.
More than 119 people did.
The project entered a community design phase. Supporters contributed design ideas. The design phase was completed. Then the app was generated and released.
For a brief moment, it looked as though the idea had finally become real.
But when people began using it, the limitation became obvious.
One supporter expressed exactly what I had hoped the app would do: help people assess accessibility using applicable requirements and comment meaningfully on levels of compliance.
What we received was essentially a generic location-and-rating application.
I wrote to App Movement and explained that the result was not functional for the purpose I had intended.
Newcastle University explained the limitation clearly: App Movement relied on predetermined templates and did not have the capacity to build tailored applications.
Users could make comments.
But the platform could not implement the structured accessibility assessment system I had imagined.
That experience crystallised something:
The idea was not impossible. The technology available to me could not yet build it.
I already had the audit methodology.
I had the individual assessment templates.
I understood the standards.
I understood lived experience.
I understood the need to connect evidence, disability, environment and compliance.
What I did not have was the ability to turn all of that into software myself.
I was not a developer.
I did not know how to code it.
And I did not have the money to employ development teams every time I had an idea.
So some ideas simply waited.
2016: Closing the Gap became BeAccessible
In 2016, Closing the Gap became BeAccessible.
The name changed, but the central purpose continued.
The employment work, Universal Design thinking, accessibility assessment, reasonable accommodation and commitment to practical inclusion continued under a new identity.
Looking back, I think of that transition this way:
Closing the Gap described the problem. BeAccessible described the direction.
Before The Door That Would Not Open
All of this happened before The Door That Would Not Open came to life.
But in another sense, that door had been present throughout my work.
Sometimes it was a physical door.
Sometimes it was the door into employment.
Sometimes education.
Sometimes technology.
Sometimes policy.
Sometimes money.
Sometimes the door was technically open, but the system on the other side was not designed for you.
That is one of the lessons accessibility has taught me repeatedly:
Exclusion is rarely one thing.
A ramp does not solve an inaccessible recruitment process.
A policy does not fix inaccessible technology.
A job vacancy does not help someone who cannot access the system required to apply.
And an app does not solve accessibility simply because it allows somebody to give a place a rating.
The door itself is only one part of the system.
Late 2025: AI changed the equation
In late 2025, I began AI training with CenteIA.
And something changed.
For the first time, AI stopped being an abstract technological development and became something I could work with.
I began developing.
I discovered that I could describe a problem, define the users, explain the workflow and work iteratively with AI to turn a concept into something functional.
The conversation no longer had to end with:
"Now I need to find a developer."
It could continue with:
"Now let us build it."
That was transformative.
Early 2026: the circle closed
When I learned how to create my own GPT, I had an almost unlimited choice of what to build first.
I went back to accessibility.
The first GPT I created was an Accessibility Audit Assistant covering the EU, US and South Africa.
I used the accessibility audit methodology I had developed years earlier, the individual templates that had lived in folders on my computer, and the standards and regulatory knowledge that had informed my work.
The first tool was built using a free tool.
That matters.
For years, one of the barriers between my ideas and implementation had been money. A bespoke developer could have cost me thousands of rands. AI reduced that cost of entry to almost nothing.
I had spent decades building the knowledge. AI finally gave that knowledge an interface.
I had been in school for decades
Looking back, I can almost say that I had been in school ever since I became disabled.
I learned disability because I had to live it.
I learned accessibility because inaccessible environments affected whether I could participate.
I learned employment inclusion because I watched capable disabled people remain outside the labour market.
I learned reasonable accommodation because employment policy without accommodation did not work.
I learned Universal Design because fixing barriers one person at a time was never going to be enough.
I learned legislation because rights without implementation changed too little.
I learned auditing because organisations needed evidence of what was wrong and guidance about what to change.
I learned technology because I could see its potential to level the playing field.
I studied formally too.
But much of my education came from decades of trying to answer one question:
What actually has to change for inclusion to work?
Then I discovered AI.
And suddenly all those years of learning had somewhere to go.
The GPTs tell their own story
Once the barrier to building came down, the ideas kept coming.
Accessibility auditing.
Digital accessibility.
Inclusive design.
Employment inclusion.
Universal Design for Learning.
HR policy.
Accessibility risk.
Bias.
Inclusive AI.
The technology changed.
The subject did not.
Looking across the GPTs I subsequently created, I can see my own history reflected back at me.
AI did not give me a new identity.
It enabled the identity and body of knowledge I already had.
AI did not make me an accessibility innovator. AI gave the innovator the tools.
When the ideas could finally move
AI also changed how I could communicate.
Through my training, I learned how to turn an idea into a moving picture.
One of the concepts I visualised was simple:
Exclusion is Designed.
Video description: The short video begins with a built environment dominated by steps and the words “Exclusion is Designed.” It then transitions to an inclusive arrangement incorporating a ramp, handrails and tactile surface cues, illustrating how different design decisions can create access rather than exclusion.
For me, that moving image captured decades of work in seconds.
Accessibility is not an afterthought added to exclusion. It is the result of designing differently from the beginning.
AI was becoming not only a development tool, but a communication medium.
It helped me build, visualise and explain.
AI did not give me the dreams
This is the part of the story that matters most to me.
AI did not give me decades of disability knowledge.
It did not create the Universal Design work.
It did not create the employment model.
It did not invent the three-legged pot.
It did not develop the original audit methodology.
It did not think of the accessibility app in 2015.
AI did not give me the dreams. It gave me the means.
It dramatically shortened the distance between something I could imagine and something I could begin building.
Some of the products I am developing today may appear new because the technology is new.
But their intellectual roots are not.
The employment intelligence work reaches back to the digital employment system I was imagining in the 1990s.
The accessibility intelligence work reaches back to the audit tools I developed and used in the early 2000s.
The modern digital assessment work reaches back to an app I tried to build in 2015.
The modular approach existed in the separate audit templates.
And the ecosystem approach was already there in my three-legged pot.
AI has helped me recognise the architecture in my own thinking.
And, finally, it has helped me build.
Maybe I was not too early
For years, I have thought:
I am always before my time.
Perhaps sometimes I was.
But today I think about it slightly differently.
Sometimes people understand a problem before technology is capable of implementing the solution they can already imagine.
That does not make the early work irrelevant.
The old database matters.
The paper checklist matters.
The website brief matters.
The audit report matters.
The failed app matters.
The three-legged pot matters.
Every one of them became part of what could eventually be built.
When the technology finally arrives, the person who has spent decades understanding the problem is not beginning from zero.
They are ready.
The circle closes
The first accessibility checklist became an audit methodology.
The methodology became a business service.
The service became part of the three-legged pot.
The three-legged pot became BeAccessible.
The methodology tried to become an app.
The technology could not yet carry it.
The idea waited.
Then I learned AI.
And the first thing I built was the accessibility tool again.
I used to think I was always before my time. Now I think perhaps I was preparing for the moment when the tools would finally catch up.
Why this matters beyond my own story
There are people everywhere carrying ideas they cannot build.
Some cannot code.
Some cannot afford developers.
Some are excluded by geography, disability, education or capital.
Some simply do not know where to begin.
I recognise that person.
For many years, I was that person too.
AI has the potential to change that relationship.
Not because everybody suddenly becomes a software engineer, but because more people can participate in turning knowledge into something usable.
AI did not begin my innovation journey. It unlocked a journey that had already been underway for nearly three decades.
For me, AI reduced dependence on specialists, development capital and technical gatekeepers.
It gave decades of accumulated knowledge a way to become products, tools, visual stories and systems that other people could actually use.
AI did not create my dreams. It made many of them buildable.
And somewhere, somebody is still standing in front of a door that will not open.
That is why I keep building.
About Fadila
Fadila Lagadien, MBA is the Founder of BeAccessible and an AI-Driven Accessibility & Disability Inclusion Innovator. Her work across disability inclusion, Universal Design, employment, accessibility and technology spans nearly three decades. Today, she explores how AI can turn deep domain expertise and lived experience into practical, accessible systems and tools.
Accessibility is not charity. It is business intelligence.
Have an inclusion problem that technology might help solve?
BeAccessible works with organisations to reduce accessibility risk, improve inclusion and design services that more people can use.
Enquire about working with BeAccessible
[executed on device: FadilaLagadien (a6080a03-d70a-4925-9730-a5688f3f9344)]