-------------------------------------------------------------------------------
PLAN:
-------------------------------------------------------------------------------
1) Use "seed" companies to discover similar businesses
• Since we already have examples (<redacted: list of customer's customers>), we will use them as “seed” companies.
• For each seed company, examine their domain and use find_similar_companies_exa(domain, num_results) to find companies that match or resemble their business model.
2) Gather a broad list of potential subscription-based companies
• Perform robust Google queries (search_google) with keywords such as:
- “subscription-based ecommerce companies”
- “direct-to-consumer subscriptions physical products”
- “DTC subscription box companies”
• Similarly, use search_exa(query, num_results) to find more companies that mention “subscription box,” “subscription service,” and other related terms.
3) Capture company details to confirm they meet the criteria
• From search results, collect potential company LinkedIn URLs.
• For each LinkedIn URL, use get_company_info(company_linkedin_url) to retrieve company size, industry, headquarters, and other metadata.
4) Filter companies based on the project’s criteria
a) Must have a subscription-based revenue model
• Verify by checking company descriptions or known business type (e.g., listing or press mentions of subscriptions, repeated purchases, etc.).
b) Must sell physical products
• Confirm via company website or LinkedIn descriptions indicating product shipments, physical goods, etc.
c) Must be direct-to-consumer (DTC)
• Confirm from descriptions, press news, or LinkedIn.
d) Must have 50+ employees
• Based on data from get_company_info to confirm current LinkedIn-reported headcount is ≥50.
5) Expand from each valid lead
• For every validated company that meets the criteria, again use find_similar_companies_exa with its domain to discover more similar businesses.
• Continue until you reach or exceed a pool of 500 potential companies.
6) Final verification
• Re-check any questionable records for headcount or DTC/subscription business model to ensure they meet the 50+ employee threshold and that they truly offer a subscription for physical goods.
7) Write the results to a CSV
• Once you have a final list of valid companies, call write_to_csv(companies) to save them, providing relevant fields such as:
- Company name
- Domain/website
- LinkedIn URL
- Employee count
- Validated subscription-based indicator
8) Print the information and confirm completion
• After writing to CSV, print or display the final list of companies (or a summary) for review.
• Finally, call instructions_complete() once all steps are done and the CSV is saved.
Post #1108
5.11K
EDU По мотивам курса про o1, о котором рассказывал на последней встрече по Gen AI, попробовал сегодня o1 модель для планирования, а потом 4o для выполнения шагов этого плана на задаче для одного из клиентов - был приятно впечатлён результатом! Глобально, с o…
Не перестаю восхищаться этим подходом с o1 - планирует, а 4o/mini - выполняет план. На днях еще попробую и DeepSeek R1 для этого, но взгляните на план по поиску subscription ecommerce бизнесов для одного из клиентов:
- 👍 23
- 🔥 7
- ❤ 6