Progressive Disclosure for AI Explanations: How Insightly Reduced Support Load

AI · 5 min read

Insightly's AI recommendations were initially delivered with verbose modal explanations that many users ignored, and a subset found them confusing or alarmist. Support tickets requesting clarifications skyrocketed, and product trust measurements dipped. Designers realized users wanted three things: a quick signal about reliability, an explanation when they asked for it, and a simple way to correct errors.

The team implemented progressive disclosure: a micro confidence badge next to each AI suggestion, an inline one-sentence rationale on hover, and a 'Why this suggestion?' expandable panel with provenance and source snippets. They also added lightweight correction tools that allowed users to retract or adjust the AI's inputs, feeding supervised examples back into the model pipeline.

Post-launch metrics showed a 42% drop in tickets related to AI output confusion, a 12% increase in users engaging with explanations (indicating better discoverability), and quicker resolution times for problems that did surface. Insightly's approach demonstrates that layering explanations—from minimal signals to full provenance—keeps interfaces clean while providing the transparency needed for user trust.