Conversational UX on a zero-dollar budget
How a scrappy, interactive FAQ proved the value of conversational design and secured funding for a fully scaled automated support ecosystem.
Chat Volume
Reduction in chat contacts
Total Volume
Reduction in overall contacts
Cost Savings
Saved per deflected chat
The Constraint
Users were drowning in a dense, static FAQ section, generating a flood of easily avoidable support tickets. A conversational chatbot was the obvious answer, but there was zero budget and no engineering resources available for a platform overhaul.
So I didn't wait. I designed a "Trojan Horse": an interactive FAQ built on the insight that e-commerce customer journeys are highly predictable. We didn't need machine learning. We needed a relentless focus on information architecture. By manually mapping existing static content into a guided decision tree, I simulated a dynamic conversational experience at virtually no cost.
The Stakeholder Bridge
To align on complex behaviors like "Close Chat" triggers, I moved away from abstract discussions. I utilized this "choose your adventure" wireframe model to socialize options directly with Customer Support stakeholders.
It gave stakeholders something to react to rather than abstract. They could feel the decision logic in real-time, which moved the conversation from subjective debate to concrete tradeoffs. That was the catalyst for final consensus.
“Our online FAQ has been fantastic for our site... Each chat contact costs us $5.00, so a reduction that significant is huge. And it’s seen as a shiny, cheap feature!”
The Scale
The success of the scrappy Interactive FAQ gave me the leverage I needed. Leadership saw the value of conversational UX, and I was granted the budget to design and build a custom, fully integrated chatbot in-house.
While the network of responses and options was sizing up considerably, the golden rule of the UX remained unchanged: never leave the user stranded. If the automated logic flow wasn’t enough to resolve the issue, I designed the system to ensure the path to a human agent was always seamless, highly visible, and instantly accessible.
Key Takeaways
- • A rough thing that works beats a polished thing that doesn't exist. The interactive FAQ wasn't pretty, but it proved the concept, moved stakeholders, and unlocked real budget. I'll always look for that scrappy first move.
- • Making stakeholders feel the logic is more powerful than explaining it. The "choose your adventure" wireframe wasn't just a design artifact. It was a facilitation tool. Putting decision logic in someone's hands changes the conversation entirely.
- • The exit ramp matters as much as the flow. No matter how good the automation gets, users will hit a wall. Designing a seamless path to a human agent wasn't an edge case. It was a core principle.
Strategic Horizon
- • The conversational data is a goldmine that I don't think is fully used yet. Every dead end and repeated question is a signal about where the flow breaks down. I'd build a regular review cycle around that.
- • I'd want to move toward more personalized flows. Right now the logic tree treats every user the same. Someone who just placed an order has a completely different set of needs than someone troubleshooting a return.
- • The long-term play is agentic support: a system that doesn't just answer questions but anticipates them based on account context. That's where this is headed, and the foundation we built makes it possible.
See the experience in action
Explore the automated support assistant flow designed to modernize AE’s support experience and build the foundation for the agentic future.