US Brand Strategy Firm · Confidential · 2024

Segment Analyzer

An experimental AI product built for a US strategy client. Type a location and an industry sector — the tool scrapes that location, gathers interior photos, runs them through GPT-4 Vision to label features and surface percentage breakdowns, then lets you compare industries head-to-head on interior-design trends. Replaced weeks of in-person fieldwork with minutes of automated analysis.

ClientUS Brand Strategy Firm · Confidential
Year2024
DisciplineExperimental Product · AI Vision · Market Research
SERPAPI · GPT-4 VISION · BATCH · USAEXPERIMENTALUSA · LOCATIONSLOCATION + SECTOR INPUTSCRAPEINTERIORSS3 · BATCHGPT-4 VISIONp-limit · BATCHLABELLABELS · LOCATION AWarm palette72%Pendant lighting61%Exposed brick48%Communal seating37%INDUSTRY COMPAREA · CAFEB · RETAILC · BANKTRENDS · GUIDANCEWEEKS OF FIELDWORK → MINUTESPROJECT · ANALYZER19
About this project

The brief, the build, and the result.

Segment Analyzer is the experimental AI-powered market-research product Remiam built for a US strategy client in August 2024. The brief: replace the slow, expensive process of sending people into individual locations to assess interior design and competitive positioning. The tool takes a location and an industry sector as input, scrapes that area for relevant venues using SerpAPI, pulls down public-facing interior photography, batches it through GPT-4 Vision (rate-limited with p-limit so OpenAI's quotas stay healthy), and produces structured labels with percentage breakdowns per location — colour palette, material use, lighting style, signage, layout, brand cues. Then a comparison view lets the client put industries side-by-side and see the patterns that distinguish them, surface trends across cities, and generate guidance for new builds. Deployed across America and used to short-cut weeks of fieldwork into minutes of analysis. Built lean as an experimental product to prove the concept end-to-end.

Highlights
  1. 01Input a location + industry — the rest is automated
  2. 02SerpAPI for location scraping and venue discovery
  3. 03Public-facing interior photography pulled and stored in S3
  4. 04GPT-4 Vision generates structured labels per image
  5. 05p-limit for safe, rate-bounded batch processing of OpenAI calls
  6. 06Percentage breakdowns per location (palette, materials, lighting, signage, layout)
  7. 07Industry-vs-industry comparison view for trend surfacing
  8. 08Replaced weeks of in-person fieldwork with minutes of analysis
  9. 09Deployed across America for a US strategy client

Build something like this with Remiam.