Cognitive Inclusion in UX Research Unlocks Better Design

Roughly 15–20% of any given population experiences some form of cognitive disability — dyslexia, ADHD, memory processing differences, acquired brain injuries. In Southeast Asia, where digital products routinely serve audiences spanning five languages, three generations, and wildly varied digital literacy levels, that number isn’t a footnote. It’s a substantial chunk of your addressable market. And yet most UX research teams treat cognitive inclusion as a compliance checkbox rather than a signal source.

That framing is expensive. Research published by Smashing Magazine from an exploratory study led by Kate Kalcevich shows that participants with cognitive disabilities consistently surface friction points that neurotypical test participants rationalise past or work around without flagging. In other words, the users most likely to abandon your funnel are the ones least represented in your research sessions.

The Research Gap That Costs You Conversions

Standard usability testing tends to recruit for availability and articulacy — participants who can show up, follow a protocol, and narrate their experience fluently. That selection bias quietly skews findings toward users who are already good at adapting to bad interfaces.

Kalcevich’s study found that participants with cognitive disabilities identified specific, actionable UX failures — unclear error messages, inconsistent navigation labelling, cognitive load spikes at form entry — that standard panels had rated as acceptable. These aren’t edge-case complaints. Error message clarity and form abandonment are two of the most well-documented conversion killers in e-commerce. On platforms like Shopee or Lazada, where checkout flows involve address validation, promo code entry, and payment method selection in rapid sequence, any unnecessary cognitive load directly impacts completed transactions.

The practical implication: if your research panel doesn’t include cognitively diverse participants, your product is being optimised for the wrong failure modes.

What Inclusive Research Actually Looks Like in Practice

Cognitive inclusion in research isn’t simply about recruiting differently — it requires adapting your methodology. Kalcevich’s findings point to several concrete adjustments:

Session structure: Shorter sessions (45 minutes maximum) with explicit transition signals reduce cognitive fatigue and produce more reliable feedback. For teams running moderated testing in Bangkok or Manila, this also maps well to local participant comfort with structured, time-bounded interactions.

Task framing: Plain-language task briefs — written at roughly a Grade 6 reading level — reduce the risk of participants struggling with the research instrument rather than the product itself. This has an inadvertent benefit: it forces researchers to clarify what they’re actually testing.

Multiple response modalities: Offering participants the option to draw, point, or describe verbally rather than type reduces the barrier for participants with dyslexia or processing differences, and often produces richer qualitative data for everyone.

The resource overhead is real but bounded. Recruiting through disability organisations in Singapore, Jakarta, or Kuala Lumpur typically adds two to three weeks to a research sprint. The payoff — a more complete picture of where your interface breaks — tends to surface in the first session.

---

AI Personalisation and the Cognitive Load Trade-off

There’s a tension worth naming here. As UX teams increasingly instrument their products with AI-driven personalisation — adaptive interfaces, predictive form completion, contextual content sequencing — the interfaces themselves become less predictable. A recent piece on UX Collective by Zeeshan Khalid raises a version of this concern from a data privacy angle: the more behavioural data AI consumes to personalise an experience, the more opaque the resulting interface logic becomes to the user.

From a cognitive inclusion standpoint, this is a specific design risk. Adaptive interfaces that rearrange navigation elements, surface different CTAs based on inferred intent, or alter information hierarchy based on session behaviour can be deeply disorienting for users with memory or processing differences — precisely the users whose experience is already under-researched.

The design principle that resolves this tension is predictable personalisation: AI systems that adapt content within a stable structural frame, rather than restructuring the interface itself. Grab does this reasonably well — promotional content changes dynamically, but the core navigation and transaction flow remain consistent across sessions. That consistency isn’t just brand discipline; it’s a meaningful accessibility feature that also reduces cognitive load for first-time users, elderly users, and anyone navigating in their second language.

For teams building on LINE’s mini-app ecosystem or designing within Shopee’s seller storefronts, where structural control is constrained anyway, this principle translates into a clear brief: personalise the content layer, protect the navigation layer.

Building the Internal Case for Inclusive Research

The barrier to cognitive inclusion in UX research is rarely philosophical — it’s organisational. Research timelines are tight, recruitment is already difficult, and the ask to broaden participant criteria can feel like scope creep to a product manager under sprint pressure.

The most effective way to reframe this is to connect inclusive findings directly to metrics the business already tracks. If your research uncovers that an unclear error message at payment confirmation is causing a 12% drop-off — and that finding came from a participant with a processing difference rather than a neurotypical tester — the fix is worth the same amount regardless of who surfaced it.

Kalcevich’s research suggests treating cognitive inclusion not as a separate accessibility workstream, but as an enrichment layer on existing research programmes. One or two cognitively diverse participants per round of testing, recruited consistently over time, builds institutional knowledge about where your product creates unnecessary friction — and creates a defensible evidence base for design investment.

For teams operating across multiple Southeast Asian markets, there’s an additional return: the interface clarity improvements that emerge from cognitive inclusion research also tend to perform better in multilingual contexts, where users are processing content in a second or third language. Simpler navigation labels, cleaner error states, and reduced cognitive load at decision points are, almost by definition, better interfaces for everyone.

---

Key Takeaways

• Recruiting participants with cognitive disabilities into standard usability research uncovers conversion-critical friction points that neurotypical panels consistently miss or rationalise past.
• AI-driven personalisation should adapt content within a stable structural frame — rearranging navigation architecture increases cognitive load for users who are already under-served by most interfaces.
• The business case for cognitive inclusion doesn’t require a separate accessibility budget — reframe inclusive findings as conversion data and connect them to metrics product teams already own.

The deeper question for design teams isn’t whether to include cognitively diverse users in research — the evidence on that is fairly settled. It’s whether the organisations running those teams are structured to act on what inclusive research reveals. Findings that challenge established design patterns require someone with enough seniority to redirect a sprint. How many research insights die in a Confluence page because no one had the authority — or the inclination — to act on them?

---

At grzzly, we help brands across Southeast Asia build research and design practices that generate commercially useful signal — not just validation for decisions already made. If your current UX process is optimised for speed over accuracy, we’d be interested in showing you what a more complete picture of your users looks like. Let’s talk