AI Virtual Assistant Development

I’m ready to move from a collection of manual workflows to a single AI-powered virtual assistant that lives wherever my users already are. The core agent you’ll build must understand natural language, respond intelligently, and carry out real-world actions across three channels—email, live chat, and voice.

Here’s what I need it to do from day one:

• Schedule appointments by talking directly to Google Calendar / Outlook APIs and handling rescheduling conflicts gracefully.
• Triage, label, and draft replies for incoming email so I only see messages that truly need my attention.
• Conduct quick research on the web, summarising sources with citations and delivering concise briefs.
• Place and receive calls (VoIP or integrated telephony such as Twilio) to confirm meetings or gather information the same way a human assistant would.

I’ll provide sample dialogues, brand voice guidelines, and access credentials. You’ll architect the agent (LLM such as GPT-4 or Claude, plus vector store if you feel it’s needed), connect the communication channels, and expose a simple dashboard so I can review conversation logs and fine-tune prompts.

Acceptance is straightforward: a live demo that shows one seamless interaction per channel completing the four tasks above, followed by hand-over of clean, commented code and deployment instructions. Let me know what frameworks or platforms you prefer and any dependencies I should prepare in advance.