Smart Quotes — AI-assisted email → quote

Smart Quotes parses incoming customer emails and drafts a quotation automatically. You connect an inbox once; every matching RFQ gets a pre-populated draft quote you review and send.

Overview

The flow:

1. Connect your work inbox (Microsoft 365 or Google Workspace) via OAuth.
2. CargoSurge polls new messages on a configurable interval.
3. An AI classifier detects messages that look like RFQs (rate enquiries, quote requests, cargo specifications).
4. For each match, the platform extracts fields: customer, mode, origin/destination, cargo, target budget if mentioned.
5. A draft quote is created with those fields pre-filled. Status: draft.
6. You review on Forwarding → Smart Quotes, tweak any field, fill in pricing, and Send.

Smart Quotes doesn't send anything on your behalf — every outbound email is still manually approved by you.

When to use

• High-volume RFQ inboxes where you get 20+ enquiries a day and manually re-typing customer details is a time sink.
• Night-shift coverage — RFQs arriving outside working hours get pre-processed so the morning team starts with drafts, not raw emails.
• Repeat-customer workflows — existing customers' emails are matched by sender domain and auto-linked to their customer record.

Step-by-step — connect your inbox

1. 1
   Settings → Integrations → Email Ingestion → Connect Microsoft (or Google).
2. 2
   OAuth flow — approve read-only Mail.Read scope. CargoSurge never sends on your behalf.
3. 3
   Choose which folder(s) to scan: Inbox, RFQs, a label you apply to quote requests, etc. Scanning the whole inbox means more noise; dedicated folder or label is recommended.
4. 4
   Set the poll interval (default 5 min; as low as 1 min on Pro+ plans).
5. 5
   Save. The first scan runs within the next interval.

What the classifier extracts

From a typical RFQ email, Smart Quotes extracts:

• Customer — matched by sender email to an existing Customer record, or proposed as a new customer if unknown.
• Mode — from phrases like "by sea", "air freight", "trucking".
• Origin / Destination — cities, countries, port/airport codes.
• Cargo — weight, volume, piece count, commodity description.
• Dates — requested pickup / delivery.
• Incoterm — if mentioned.
• DG flag — if the email references hazardous materials.

Anything not found stays blank for you to fill in.

Reviewing a draft

Forwarding → Smart Quotes shows the inbox of drafts:

• Each row shows the source email subject, the extracted fields, and a confidence score.
• Click a row to open the draft quote side-by-side with the original email.
• Edit any extracted field — your corrections train the classifier for next time.
• Fill in charges (the one thing AI can't guess — it's your pricing).
• Approve as Quote — moves the draft into the regular Quotations list with status draft.
• Discard — if the email isn't really an RFQ, skip. Also trains the classifier.

Common pitfalls

Privacy & security

• CargoSurge only reads emails, never sends.
• Emails are stored only long enough to extract fields; the raw body is purged after 30 days.
• Your OAuth token is encrypted at rest and can be revoked anytime from Integrations.
• PII inside emails (customer contact details) is treated per your tenant's data-retention config.

Limitations

• Works best with English-language RFQs. Arabic and other languages have lower extraction accuracy today.
• Very short emails ("can you quote Dubai to Mumbai?") produce mostly-empty drafts — the AI won't hallucinate cargo specs.
• Forwarded email chains are parsed but can misattribute the original sender; double-check the customer match.

Turning it off

Settings → Integrations → Email Ingestion → Disconnect. Existing drafts remain; no new drafts are created.