AI-assisted development is transforming how quickly we build software. Around 90% of developers now use AI tools in their daily work¹, but AI-generated code does not automatically produce accessible output²,³. This creates a growing risk: as development accelerates, accessibility can fall further behind unless teams apply practical interventions, such as building guidance into the tools and workflows they already use. For the 1.3 billion people worldwide who depend on accessible digital experiences⁴, that gap matters — and industry data shows proactive accessibility adoption has plateaued despite growing awareness⁵.
That widening gap points to a clear imperative: accessibility must be taken as a priority earlier in the software development lifecycle, before issues are created and scaled.
Why accessibility needs to shift left
The solution is to build in accessibility from the start. This means embedding accessibility guidance directly into development workflows and meeting teams where they already work. It combines human expertise with AI-assisted validation, focusing on prevention during design and development rather than detection after the fact.
McKinsey research confirms what many teams are discovering: embedding accessibility earlier in the software development lifecycle reduces costly rework and improves overall quality⁶. When the right approach is applied at the right time, even complex challenges become more easily addressed. Most accessibility issues follow predictable patterns — and when we understand those patterns, we can turn them into proactive guidance rather than reactive fixes.
SAP’s approach to accessibility in AI-assisted development
At SAP, we’re responding by exploring how accessibility guidance can be brought closer to the point of creation in AI-assisted development. Our focus is on helping teams consider accessibility earlier, apply established guidance more consistently, and keep expert oversight central to the process.
This work is grounded in a simple principle: AI can help make accessibility guidance more available in everyday development workflows, but it should not replace human judgment.
Impact beyond standards
This work goes deeper than accessibility standards alone. By embedding accessibility guidance at the point of creation, we’re making accessibility part of how teams design, build, and review software every day. That helps teams catch recurring patterns earlier and reduce manual review work in later phases. The impact is both technical and human. For the 1.3 billion people worldwide who depend on accessible digital experiences⁴, this approach means products built with their needs in mind from the very first line of code. It means fewer barriers, better experiences, and software that works for everyone.
This work is part of a broader industry movement toward accessible AI-assisted development — one that will require organizations to share what they learn, align around practical standards, and keep accessibility visible as AI changes how software is built.
Join the movement
SAP is committed to making accessibility a core part of how we build with AI. Whether you’re exploring similar approaches in your organization, embedding accessibility earlier in your development process, or looking for practical ways to combine AI speed with inclusive design, we invite you to explore our resources and join the conversation:
- Accessibility and Inclusive Design — Learn more about SAP’s approach to designing inclusive experiences.
- Building Empathy into Design — Explore how SAP is redefining accessibility through empathy.
- Accessibility Design Tools Figma Library — Explore SAP’s Figma Library for embedding accessibility into your design workflow.
References
¹ DeBellis, D., Storer, K. M., Harvey, N., Beane, M., Edwards, R., Fraser, E., Good, B., Kalliamvakou, E., Kim, G., Maxwell, E., D’Angelo, S., Murillo, A., Inman, S., & Villalba, D. (2025). State of AI-assisted software development 2025. DORA; Google Cloud.
² Gurita, A-E., & Vatavu, R-D. (2025). When LLM-generated code perpetuates user interface accessibility barriers, how can we break the cycle? Proceedings of the 22nd International Web for All Conference, W4A '25, 124-134.
³ Panchanadikar, R., Bhosekar, M. S., & Dixon, E. (2025). Can generative AI create accessible websites? Proceedings of the 27th International ACM SIGACCESS Conference on Computers and Accessibility, ASSETS '25, 1–6.
⁴ World Health Organization. (2023, March 7). Disability.
⁵ Level Access. (2025). Seventh annual state of digital accessibility report 2025-2026.
⁶ McKinsey & Company. (2025, February 10). How an AI-enabled software product development life cycle will fuel innovation.
