Multi Purpose AI Agent Development

I have a small in-house codebase that already runs a few large-language-model helpers, but I now need to expand it with three new agents: a Sales Assistant, an Email Agent, and a Document Management Agent. All three must plug into the same Python stack (LangChain, OpenAI API, vector store, webhook layer) and respect our existing coding patterns.

Here is the core scope we will iterate on together:

• Sales Assistant Agent
– Lead generation from incoming data feeds and web sources
– Follow-up communication that can draft, personalise, and schedule messages
– Basic sales analytics so my team can see pipeline health at a glance

• Email Agent
– Automated responses that pick the right template and tone
– Sorting and categorisation into predefined folders/labels
– Content analysis to flag intent, urgency, or sentiment
– Ability to read attached documents (e.g. purchase invoices) and convert key data into structured JSON for downstream systems

• Document Management Agent
– Ingest and index new files into our vector database
– Provide conversational retrieval over stored documents
– Enforce role-based access using the permissions model we already have in place

Success criteria: each agent runs as an independent service with a simple REST endpoint, ships with unit tests, and is delivered as a clean, well-documented Git repository that my developers can pull and deploy. I will share API keys, sample datasets, and further business logic once we open a chat here on Freelancer. If you have recent experience building production-ready agents on top of GPT-4, LangChain, or similar frameworks, I’d love to review your approach and timeline.