Local-first AI in Telegram to curb mass Web3 outreach

Mass automated outreach and manual CRM workarounds are harming partnerships in the Web3 sector, according to Sean Assaf, head of business development at Telebiz. He recommends moving CRM administration into Telegram and using local-first AI to keep conversation content private and preserve human control over outreach decisions.

Deal teams commonly conduct conversations in Telegram and then copy details into separate CRM dashboards. That two-step routine can create gaps in pipeline data and expose chat content to third-party tools. Telebiz runs as a browser client that connects directly to Telegram’s servers. When an AI feature reads a thread for context, requests go from the user’s browser to the model provider the user chooses. Telebiz retains only an encrypted business layer — chat IDs, mappings to HubSpot deals, follow-up rules and timestamps — with per-organization keys and does not hold raw conversation text.

Assaf identifies administrative tasks suitable for automation: retrieving threads, producing summaries, drafting message options, logging agreed terms and pursuing cold follow-ups. Outreach timing and the message content that affect trust should remain under human control. He captured the distinction with a direct guideline: “Automate everything that happens before and after a decision. Never automate the decision.”

High-volume, generic outreach has reduced trust in relationship-driven fields such as venture capital and Web3, Assaf notes. Many relevant contact lists are small and contacts often know each other, so repetitive blasts can damage relationships. He recommends using AI to detect specific external triggers — a funding round, a mainnet launch or a senior hire — and flag those moments for human engagement.

Embedding CRM functions into chat shifts routine data capture to the moment conversations occur. Assaf expects dashboards to remain for reporting and analytics while day-to-day operations move into messaging apps. He also foresees developers adding team-focused tools inside Telegram for shared context, relationship ownership and clean handoffs when staff leave, with multiple such features appearing by 2026.

He cautions against automating outbound messages at scale and recommends directing automation toward administrative efficiency and context monitoring rather than replacing human judgment.