Empathy, Accessibility and the Limits of AI Trust (John McCartney)

The podcast discusses the importance of human-centered design in software development, emphasizing empathy, inclusivity, and real-world user feedback. Observing how diverse users interact with products - especially those with disabilities - reveals critical usability issues that internal teams might overlook. The discussion highlights the need for cross-functional collaboration, where accessibility is a shared responsibility among designers, engineers, QA, and product owners, rather than siloed or treated as an afterthought. Usability testing with real users, particularly those using assistive technologies, leads to "aha moments" that drive meaningful improvements and foster organizational change.

Key themes include the value of small, iterative wins over large-scale overhauls, the limitations of AI without human oversight, and the dangers of designing based on assumptions rather than evidence. The conversation draws parallels between teamwork in hockey and software development, underscoring collective accountability, trust, and resilience. Personal stories around cancer advocacy illustrate how deeply held values can fuel sustained commitment to community-driven causes. Ultimately, the podcast advocates for a mindset shift toward curiosity, humility, and continuous learning - where quality is defined not by technical precision but by a product's ability to solve real human problems in inclusive and empathetic ways.

• What if you observed a real user with a disability interact with your software for 30 minutes this week?

  • Move: Schedule and conduct a live usability test with a person who uses assistive technology (e.g., screen reader, voice navigation), either through a community group, accessibility network, or freelance platform like UserTesting with accessibility filters. Focus on one core feature of your product.
  • Why Now?: AI-generated assumptions about user behavior are increasing - without direct human feedback, your product risks reinforcing biases or missing critical interaction flaws that automated tools won't catch.
  • Expected Upside: Uncover at least one overlooked usability barrier (e.g., navigation flow, contrast issue, labeling) that you can fix immediately, improving both inclusivity and overall UX quality - potentially preventing churn among underserved users.

• What if you shipped a "bunt" version of your next feature - just one small, testable slice - in the next 72 hours?

  • Move: Break your current backlog item into micro-components and deploy the smallest possible version that delivers observable value (e.g., a single form field with validation, not the full onboarding flow). Measure how users engage with it via analytics or direct observation.
  • Why Now?: Waterfall-style planning increases risk; solo developers can leverage speed to validate ideas before over-investing. The market shifts fast - iterative delivery lets you adapt without burnout.
  • Expected Upside: Gain real-world feedback within days instead of weeks, reduce rework by up to 50%, and build momentum by shipping something tangible - proving progress compounds faster than perfection.

• What if you made yourself redundant in one part of your workflow by documenting and automating it this week?

  • Move: Identify a repetitive task (e.g., deployment, testing, customer onboarding) and create a self-running system: write a script, set up a CI/CD pipeline, or build a no-code bot with clear instructions so anyone (or any future version of you) can operate it without explanation.
  • Why Now?: As a solo operator, your leverage comes from systems - not just coding. With AI accelerating tooling, the window to embed human-centered logic into automated workflows is now - before defaulting to generic solutions.
  • Expected Upside: Free up 3 - 5 hours per week within a month, reduce errors, and create a foundation for scaling trust in your product - aligning with the principle that quality means solving real problems reliably, not just shipping code.

• Conduct regular usability testing sessions with real users, especially those with diverse abilities, to uncover hidden accessibility issues and validate design assumptions early.
• Implement a shared team ownership model for accessibility by defining clear roles (design, engineering, QA, product) in the development lifecycle to ensure inclusive outcomes.
• Prioritize small, iterative improvements ("bunts and singles") over large-scale changes to build momentum, demonstrate value, and reduce risk in product development.
• Replace hypothetical personas with proto-personas based on real user research, particularly including people with disabilities, to ground design decisions in actual user needs.
• Integrate direct user feedback - especially from underserved groups - into sprint cycles to close empathy gaps and drive meaningful, evidence-based product iterations.