Decisions should start with users, end goals, and the whole flow. Like web design, fewer clicks don't always make a path easier to follow — and on a phone call, people often need more context to answer with confidence. That can run directly against the business goal of shorter handle times, so the work is knowing when to cut and when to add.
The hardest part of designing for legacy IVR systems is holding two things at once: real technical constraints, and the case for users even when handle time or agent-routing metrics don't move. Our job is finding the balance between profitability, which pushes for fewer transfers to agents, and the frustration people feel talking to a machine.
The current industry trends are moving away from valuing UX research as a critical superpower in a designer's tool belt. With budgets and project timelines being cut as lean as possible to catch up to industry peers, the argument is: why do we need traditional research methods such as interviews, usability tests and in depth heuristic analyses when we can just ask AI to scour the internet and provide all the main insights we need?
This is the biggest challenge I've faced as a designer working in conversational AI. My belief is that AI tools can indeed be used to support smaller teams and hard deadlines to provide proxy research as reference for design decisions. However, there is constantly a risk of AI hallucination (spouting statistics and data where there is no real data) and with the new technology available that can give publishing and media sites tools to block AI from scanning their information, there is always a possibility of unreliable references affecting your million dollar project.
Call me old-school, but I love nothing better than to discover new (or perhaps validate existing) insights through real, human dialogue and conversation. My core passion in designing for conversational design is not to replace human interactions, but to provide fast, simple and straightforward outlets that can guide users to their end goals. I advocate best for design decisions rooted in real data, and believe there's a place for AI as well as UX research to have its place to shine.
Lucid
Visio
Genesys
Figma, Figma Make
Google Gemini
Writer
Anthropic Claude
Active collaborations with Content Strategists and Writers
Lots of coffee
A sense of humor when the testing goes wrong
