24/7 Customer Service via AI Chatbot

The problem worth solving

The company's customer service was entirely human-staffed — meaning support was unavailable outside business hours, and agents were spending significant time handling repetitive, low-complexity queries that didn't require specialist knowledge. As the customer base grew, this model wasn't scaling. After-hours contacts went unanswered, first-contact resolution rates were constrained by queue times, and agent capacity was being consumed by queries that a well-designed automated system could handle.

What I did

From zero to launched — a discovery-led approach to the company's first AI product.

• User journey mapping: Mapped the customer service user journey end-to-end — identifying the query types, touchpoints, and escalation paths that would determine what the chatbot needed to handle versus route to a human agent.
• FAQ set creation: Built out a comprehensive set of FAQ content to power the chatbot's responses — working with the CS team to ensure accuracy, tone, and completeness across the highest-volume query categories.
• QA / evaluation framework: Designed and built the quality assurance and evaluation framework in collaboration with CS and technology teams — defining how chatbot responses would be assessed for accuracy, relevance, and appropriate escalation behaviour before and after launch.

The outcome

The company's first AI chatbot launched successfully — providing 24/7 customer service coverage for the first time. Customers received immediate responses to routine queries outside business hours, and human agents reported meaningful reduction in repetitive contact volume. The QA and evaluation framework ensured the chatbot's responses met quality and accuracy standards before going live — and continued to provide a structured feedback loop for ongoing improvement post-launch.