All use cases

01Case studyCult.fit

Cult.fit. When AI didn’t know your stores existed.

17 stores. Four AI assistants. One question: when a Bangalore shopper asks ChatGPT for a fitness store nearby, does Cult come up? Shopthru AI measured the answer across 1,104 local questions.

Industry
Local commerce
Headline outcome
6.6% discovery
01Measures

Measured every kind of local question, every store, every assistant.

The team asked 1,104 local questions about 17 stores in Bangalore and Hyderabad across OpenAI, Perplexity, Copilot and Gemini, from questions that name the brand to discovery questions that do not ("stores like X near me"). Questions that named Cult were recognized 90 to 100% of the time. Discovery questions collapsed to single digits.

  • 6.6%found when shoppers asked for a store nearby
  • 293hallucinated hours / addresses caught
  • 1,190unique competitors mapped
02Simulates

Checked the facts: the live store listings became the source of truth.

Every store fact AI quoted was checked against the stores’ live public listings. The team surfaced 86 wrong store hours, 101 wrong addresses, 14 wrong phone numbers, and 29 wrong ratings, including a Hyderabad location where Gemini answered "open until 10:30 PM" for a store that closes at 9:00.

  • 4,943 AI answers read, duplicates removed, every store fact checked against the listings
  • 789 rival websites mapped in one pass
  • 1,366 rival name spellings matched to 1,190 distinct rival brands
03Ships

Shipped the fix queue: every local gap became a top-priority task.

The team opened 318 tasks to get Cult listed on the local sites AI reads (Justdial, magicpin, Shopsy, YouTube, Instagram) and 130 tasks to close gaps with rivals, area by area. It queued 91 more for the stores themselves: too few Google reviews, missing local schema, missing area pages and missing trust signals.

  • 318 local listing tasks: the third-party sites the brand didn’t own
  • 130 rival gap tasks: area-by-area parity work
  • 91 AI search tasks: schema, Google Business Profile reviews and area pages
04Compounds

Compounded into brand memory: the local question scoreboard now runs for every brand with stores.

The full set of local questions and the store listing check first ran for Cult. Both are now defaults for every brand with physical stores the team onboards. The "AI invents store hours" pattern alone has been seen in all three such brands audited since, and it now flags automatically in every new local discovery scan.

See what AI would buy from you.