Skip to main content
Reading hundreds of customer reviews manually takes hours and still misses patterns. Teams end up guessing what to fix instead of knowing. This prompt turns a raw batch of reviews into a structured analysis — sentiment breakdown, top pain points, feature ideas, and a priority ranking — ready for your next roadmap session. Just fill in 4 inputs and get a team-ready brief in seconds.

How the prompt works

The system prompt sets the LLM up as a product analyst and requires five specific sections in every response. Without that structure, models return a vague summary instead of something a product team can act on. The user prompt takes four variables — product name, review text, focus areas, and current priorities — so the output is always scoped to what the team is actually working on.

System prompt

Sets the LLM’s role and defines the five sections it must always include in the output.

User prompt

Four variables that give the model enough context to return insights scoped to the team’s current priorities.
Customer review analysisOverall sentimentMixed sentiment — 60% negative, 40% positive. Customers appreciate product selection and deals but are frustrated with core functionality.Top 5 recurring pain points
  1. App crashes during checkout (High) — direct revenue loss, cart abandonment
  2. Confusing payment process (High) — conversion rate reduction
  3. Poor search functionality (Medium) — product discovery issues
  4. Inaccurate delivery tracking (Medium) — increased customer support load
  5. Cluttered interface design (Low) — general UX degradation
Feature improvement ideas
  • Immediate: Fix checkout stability and implement crash reporting
  • Short-term: Redesign payment flow with progress indicators
  • Medium-term: Improve search with better filtering and AI suggestions
  • Long-term: Add wishlist functionality and fix delivery tracking integration
Priority classification
  • High: Checkout stability, payment flow redesign
  • Medium: Search enhancement, delivery tracking accuracy
  • Low: UI decluttering, wishlist feature
Key quotes
  • “Love the product selection but payment process is confusing.”
  • “App crashes when I try to checkout.”
  • “Great deals but search function is terrible.”

Import into Adaline

This prompt comes with a ready-to-import Adaline project file. It includes the prompt, dataset, and evaluators, all pre-configured.

Evaluations and dataset

Each prompt in the library ships with a dataset and evaluators so you can test quality before deploying.

Evaluators

Two failure modes, three evaluators: one for completeness, one for actionability, one for length.

Output completeness

Checks that all five required sections are present in every output.

Actionability

Checks that insights are specific enough for a product team to prioritise and act on immediately.

Response length

Guards against bloated output. A good analysis is a quick brief, not a full report.
A good analysis is concise and scannable. If the output is too long, teams won’t read it.

Dataset

Four product types across different industries — each row maps directly to the four variables in the user prompt.