Accessibility & Portal Redesign.
How do you redesign a grant portal used by tens of thousands of researchers without locking anyone out of federal funding?
Executive summary
NSERC and SSHRC fund the bulk of Canadian university research. The portal where researchers, reviewers, and institutional administrators apply, peer-review, and manage grants had grown over a decade into a system most users tolerated rather than trusted. I joined as senior researcher and accessibility lead on a multi-year redesign, embedding WCAG 2.1 AA conformance into research recruitment, design decisions, and component-level acceptance criteria.
The result is a portal where accessibility is a baseline behavior, not a remediation pass at the end of the build.
The problem.
A grant portal is the gatekeeper to a research career. If a screen reader can't navigate the application form, that researcher does not file. If a peer-review interface burns 90 minutes on a 30-minute task, reviewers quietly stop accepting invitations. Both of those failures distort science funding, and neither shows up in a usability survey.
The redesign couldn't be a redesign in the cosmetic sense. It needed to become the moment two federal agencies bound accessibility, research, and design into one operating model.
My role.
UX researcher and accessibility lead, working alongside designers, content strategists, and the engineering build team.
- Recruited inclusively, by default.
Sampling included screen-reader users, low-vision users, motor-impaired users, and cognitive-diverse participants from the start, not in a separate "accessibility round" at the end.
- Wrote accessibility into acceptance.
Every component shipped against acceptance criteria covering semantics, keyboard support, focus states, contrast, and screen-reader output. The criteria sat in the same backlog as functional ones.
- Led usability testing across the full user mix.
Application flows, peer review, and administrator dashboards were each tested with the populations who actually use them, including in French and on assistive tech.
- Translated standards into design decisions.
WCAG is a target, not a design system. I worked with designers to convert the standard into specific decisions about form structure, error recovery, and time-bound interactions.
Unique challenges.
- Two agencies, one portal.
NSERC and SSHRC share infrastructure but serve different research communities with different cultural conventions. The system had to be one product without erasing either agency's identity.
- Bilingual from the protocol up.
Federal portals are simultaneously English and French. Research, content, and accessibility testing all needed to hold up in both, not be translated post-hoc.
- Legacy data, modern interface.
The portal sits on top of records dating back years. The new front end had to honour the historical record without inheriting its UX assumptions.
My process, highlights and takeaways.
- Accessibility moved upstream.
The earlier accessibility entered the conversation, the cheaper it was to honour. By the time components hit engineering, semantic structure was already in the design files.
- Peer review got the time it deserved.
The peer review interface had been an afterthought for years. We treated it as a first-class product, with its own research and dedicated component work.
- Errors as a design surface.
Most grant-portal failures aren't catastrophic, they're recoverable. Designing error messages, validation timing, and save-state behaviour explicitly turned out to be one of the highest-leverage accessibility decisions in the project.
- Component-level audit trails.
Every accessible-by-design decision was traceable to a research finding or a standards clause. That trail mattered when scope debates surfaced eighteen months in.
Final thoughts.
What worked well. Embedding accessibility into research recruitment from week one set the cultural tone for the whole engagement. By the time we got to component-level work, "is this accessible?" wasn't a separate conversation, it was already inside every other conversation.
What I would change. I would have invested earlier in cross-agency content governance. We solved the design and component problems before the content model fully caught up, which created some retrofitting late in the project.
Thanks to the NSERC and SSHRC teams who treated accessibility as an organizational commitment, not a checklist.