LLM-Emails

Deliverables by the end of March.

• Requirements & Process Flow:
  • Application Interface: An interface by which the customer can interact with.
  • Crunchbase: getting the companies data into the interface by filtering:
    • Getting JSON from Crunchbase based on customer-preferences such as:
      • Company size
      • Domain(s) of operation.
      • Personal details of C-level execs (CTO, CMO, etc..)
      • Based out of (location).
      • Series of funding (if it's a startup), etc...
    • Details of the social media profiles (main focus on LinkedIN) of the execs is taken.
  • LinkedIN: Using the linkedin profile urls of the execs, the following should be extracted:
    • Name.
    • Current position.
    • Location
    • Previous Experience.
    • Education, etc...
  • Knowledge Graph and database:
    • Using all of the data collected above, a knowledge graph should be shown in the Application.
    • The details should be stored in neo4j AuraDB.
  • Zoho CRM:
    • Using the personal details of the execs, leads should be created.
    • The mail context from LLM is used to send an email to the lead.
    • Based on the reply, the conversation is taken forward.
    • Workflow is automated to continue the conversation (with required permissions from employees, and monitoring).
    • If lead is interested, then move onto negotiations for:
      • Product requirements.
      • Pricing
      • Support, etc..
  • LLM, embeddings and vector storage:
    • The context extracted above should be stored in a database and then embeddings created.
    • The model embeddings are also to be taken.
    • A framework (like langchain) is to be used to build this RAG application and knowledge graph.
    • The LLM uses this context and then creates a hyper-personalized mail context.

• Products required (all are paid):
  • Crunchbase: for getting company and execs' personal data.
  • Zoho CRM: A CRM to manage marketing, sales and mail conversation with the leads.
  • LinkedPro: An extension to scrape, extract and obtain LinkedIN profile data of the execs and populate as leads in ZOHO.
  • OpenAI / Gemini (for LLM).
  • Costs for the products (with upfront investment). Link Pricing details of various Softwares.
• Product that isn't paid:
  • AuraDB: for storing and managing neo4j graph database in the cloud.
• Products that are optional to be used:
  • Crunchbase Databoost and apptopia (even more company data) link.

• Further questions:
  • Should LLM-generated emails be extended for logistics, inventory and financials?
  • Should the mails be monitored with the use of a email-marketing, analytics tool such as Mailchimp?