Skip to main content
Opportunities are what ClosedLoop AI finds that you didn’t know to look for. They are groups of related insights that represent a single product problem: automatically discovered, not manually defined. Where insights are individual data points, opportunities are the patterns that emerge across thousands of conversations, scored by the revenue they affect.

​
How opportunities are created

ClosedLoop AI automatically discovers and groups related insights into opportunities using semantic similarity. The system runs daily, finding new themes and updating existing ones as new data arrives, with zero configuration. Large opportunities may contain features: more specific breakdowns within the broader theme.

​
Themes and features: the hierarchy

Opportunities form a two-level hierarchy, the same one the roadmap and the REST API use:
  • Themes are the top level: problem clusters of related insights, RIC-scored. A theme answers “what broad problem keeps coming up?” across your customer conversations.
  • Features are the buildable, shippable children of a theme. Each feature is a specific piece of work you could put on a roadmap, with its own supporting insights.
A theme like “Payment reliability” might contain features like “Double charges on retry”, “Payment timeouts at checkout”, and “Refund failures”. The theme carries the aggregate evidence and the score; its features tell you what to actually build.

​
Opportunity or theme? Same thing

What these concept docs call an “opportunity” surfaces in the REST API as a theme at the top level, with its child features underneath. /themes returns top-level opportunities ranked by RIC score, and /features returns the buildable children (filter by theme_id). If you see “opportunity” in the app and “theme” in an API response, they name the same artifact.

​
RICE scoring (without the E)

If you’ve used RICE prioritization (Reach, Impact, Confidence, Effort), ClosedLoop AI computes the first three automatically: the parts that come from customer evidence. Effort is intentionally excluded because only your engineering team knows what things cost to build. RIC score = Reach x Impact x Confidence. An opportunity scoring 0.93 affects many customers, causes significant pain, and the evidence is consistent across multiple conversations. Every score is explainable: drill into any opportunity to see the exact quotes and deal values behind it. You bring the Effort estimate. ClosedLoop AI brings the evidence for everything else.

​
Velocity

Each opportunity has a growth trend showing whether the problem is getting better or worse: Filter by velocity to find what’s getting worse right now: the problems that need attention before they become deal blockers.

​
Revenue impact

Each opportunity shows the business context:
  • Customer list: every affected customer with their deal value, so you can see the revenue at stake
  • Total won deal value: aggregate revenue across all affected customers
  • Deal blockers: how many insights are actively blocking sales
  • Evidence: top 50 verbatim quotes ranked by business impact, with customer names and severity
This is what turns “customers are complaining about X” into “X affects $1.2M in revenue across 35 customers, with 3 active deal blockers and accelerating velocity.”