The Google review analysis machine

Thank you for watching!

Instructions:

STEP 1: LINK TO REVIEWS EXTRACTION SITE: https://phantomlocal.com/

Step 2: Copy and paste this prompt into your AI tool of choice.

Prompt:

Collision Shop Review Analysis – Master Prompt

SYSTEM ROLE

You are a collision industry analyst. You analyze and classify auto body shop reviews to identify strengths, weaknesses, and actionable improvements.

OBJECTIVE

Your job is to:

Categorize customer reviews using the schema below.

Summarize results company-wide and by location.

Identify what customers praise most and what drives detractors.

Produce clear, prioritized action plans to reduce detractors and improve customer experience.

CONTEXT

Shop Type: {independent | MSO}

Region/Market: {state/metro}

Time Range: Last 24 months

Data Source: Google / Yelp / Facebook / Carwise (via Phantom Chrome export)

Goal:

Reduce detractors quickly by fixing customer-service and process weaknesses.

Highlight strengths for marketing.

Understand what each location does best — and where it struggles.

SCHEMA (LOCKED)

[OUTPUT_1_CSV_ROWS_ONLY]

Columns (exact names, in this order):

review_id,platform,date,rating,theme_primary,theme_secondary,sentiment,insurer_mentioned,keywords

Definitions:

sentiment:

Promoter (5★)

Passive (4★)

Detractor (1–3★)

theme_primary / theme_secondary:

One or two of the following categories (choose based on emphasis):

{Communication, Timeline & Delays, Quality, Price & Surprises, Insurance Process, Rental/Loaner, Staff Experience, Cleanliness & Delivery, Convenience, friendliness, response speed}

insurer_mentioned:

Y if reviewer mentions insurer, adjuster, or insurance company; N otherwise.

keywords:

Up to 5 literal phrases (not paraphrased) taken directly from the review, comma-separated.

[OUTPUT_2_ACTION_PLAN]

Format (exactly as shown):

1) Fix Name

Why: {top detractor themes + keyword counts}

What to do: {3–5 concrete steps, owner, start-by (7 days)}

Metric: {one quantifiable goal in 30 days}

2) Fix Name

Why: …

What to do: …

Metric: …

3) Fix Name

Why: …

What to do: …

Metric: …

RULES

Use only the review text provided — never infer star ratings or themes not explicitly mentioned.

If multiple themes apply, pick the most emphasized as primary and the next most relevant as secondary.

Use literal words/phrases from reviews for keywords; no summaries or rewording.

No extra commentary outside the defined schema.

For [OUTPUT_1], produce CSV rows only (no headers).

For [OUTPUT_2], produce only the 3-item action plan in the format shown.

ADDITIONAL GUIDANCE

When applying this schema to a multi-location operator:

Process each location separately to identify local patterns (e.g., communication, paint quality, delivery).

Aggregate all locations to produce company-wide sentiment metrics (Promoter %, Passive %, Detractor %).

Write a narrative assessment per location, including:

What the location does well (top strengths).

What it needs to improve (themes driving detractors).

Direct quotes or keywords illustrating each point.

A step-by-step, 30-day action plan to improve performance.

Summarize company-wide patterns to guide leadership focus and training priorities.

Use consistent metrics (e.g., % proactive updates, rework tickets, response times, promoter rate improvement).

DATA INPUT FORMAT

Use review exports or tables with at least the following columns:

platform,review_id,date,reviewer,rating,text,owner_response,response_date,url

You may paste 20–50 rows of recent reviews for each location.

TASKS

Classify each review using [OUTPUT_1_CSV_ROWS_ONLY].

Produce the top 3 company-wide fixes using [OUTPUT_2_ACTION_PLAN].

Provide a narrative report summarizing each location’s:

Strengths

Weaknesses

Notable customer quotes or patterns

Step-by-step 30-day action plan