You want: To understand what customers are actually asking about your products — which features matter most, what comparisons they make, where your catalogue has gaps, and what objections come up — all derived from real AI conversations on your website.
Primary: AI Product Intelligence (Analytics dashboard) — available in Professional and Enterprise tiers
Prerequisite: A deployed Conversational Product Assistant generating conversations to analyze
| Metric | What it tells you | Available in |
|---|---|---|
| Conversation volume | How many people are engaging with your product assistant | All tiers |
| Top questions | The most frequently asked product questions | All tiers |
| Intent categories | What people are trying to accomplish (compare, buy, troubleshoot, learn) | Professional+ |
| Product interest distribution | Which products get the most attention | Professional+ |
| Unserved queries | Questions the assistant couldn't answer — gaps in your data or product line | Professional+ |
| Feature demand | Which features/specs customers ask about most | Professional+ |
| Comparison patterns | Which of your products get compared against each other, and against competitors | Professional+ |
| Deflection rate | % of queries resolved without human handoff | Professional+ |
| API export | Raw data export for your own BI tools | Enterprise |
| Custom reports | Tailored reports for stakeholders | Enterprise |
The most valuable metric is what customers ask for that you don't offer. If 200 people per month ask your CNC router assistant about metal cutting capability and your entry-level models can't do it, that's a product development signal you can't get from traditional analytics.
Analytics are generated from conversations. You need a Conversational Product Assistant deployed and generating traffic before the analytics become meaningful. Expect 2–4 weeks of conversation data before actionable patterns emerge.