The owner's problem with AI
Most AI pitches ask you to trust a black box with your client relationships. One wrong autonomous email to a top client costs more than a year of saved typing. That is why 4logist AI is draft-only on every channel — email, WhatsApp, everywhere. The agent prepares; your team sends.
The second owner problem is unpriced revenue leaking out quietly: a request sits open, nobody quotes, the client books elsewhere. The Deal Watcher nudges your Telegram when a 4logist request sits open and unquoted past your threshold — once per request, no spam.
What you get without hiring
- Overnight and weekend RFQs arrive as reviewable drafts, priced from your historical lanes — not market averages.
- Customs packets extracted with OCR grounding; unsafe packets wait for a human instead of filing a wrong party name.
- One Telegram bot per workspace for approvals, questions, and deal alerts — strangers who find the bot are refused.
- Usage-based credits: a light quote costs less than a hard one, and the ledger shows every charge.
Control model
Quotes are drafts until an operator taps Send. Document auto-filing is gated: only packets with zero alarms, clear sanctions screening, a confident order match, and high extraction confidence file on their own — and you can require a human tap on every packet with one switch.
Frequently asked questions
Can the AI email a client without approval?
No. There is no auto-send path on any channel. A quote reaches a client only when an operator sends it from the dashboard.
What does it cost when the team barely uses it?
Billing is credit-based and metered on real AI actions. Triage-rejected mail, errored runs, and unsent drafts are not charged.
Do I need to replace my TMS?
No. 4logist AI sits beside your 4logist TMS and your exchanges — it reads lane history and can file requests, but the TMS stays the system of record.