Translation
Workflow
Automation
An end-to-end AI platform that replaces your manual email coordination with a 14-stage automated pipeline. From client intake to invoice — without the overhead.
Translation agencies run on email.
That's the problem.
One operations person. Hours per day reading emails, assigning translators, chasing updates, forwarding QA reports, generating invoices. Every project manually coordinated. Every status check a new email thread.
- ×Read every client email manually
- ×Calculate word counts from attachments
- ×Email translators one by one
- ×Track deadlines in a spreadsheet
- ×Forward QA feedback manually
- ×Write and send invoices by hand
- ×No visibility into job status
- AI reads and classifies all intake emails
- Word counts extracted automatically
- Translators assigned in priority order
- Deadline reminders sent automatically
- QA reports routed with zero manual work
- Invoices generated and delivered on completion
- Real-time dashboard for every active job
14 stages. Fully automated.
Every stage triggers automatically on state change. No polling, no timers — event-driven from intake to invoice.
Client Email Intake
Stage 1AI reads incoming client emails, classifies project type, and parses Excel/PDF attachments to extract word counts automatically.
AI Classification
Stage 2GPT-4o classifies document type, language pair, urgency, and project metadata — creating a structured project record.
Project Created
Stage 3A complete project record is created with all extracted data — word count, client, document type, deadline window.
Translator Offer
Stage 4Korean-language job offers sent to translators in priority order. Each offer includes scope, rate, and deadline.
Accept / Decline Detection
Stage 5Inbox is polled every 2 minutes. On decline or timeout, the system automatically advances to the next translator.
Client Notified
Stage 6Client receives confirmation with project ID and assigned translator — all emails reviewed by ops manager before sending.
Deadline Tracking
Stage 7Active deadline monitoring across all concurrent projects. Reminder sent to translator 1 hour before due time.
Completion Detection
Stage 8Translator delivery is detected automatically. Project advances to QA stage with no manual intervention.
QA Management
Stage 9Client QA feedback emails are detected, parsed, and routed back to the translator with attachments. Revisions tracked.
Invoice Generation
Stage 10Invoice automatically generated and delivered when the project completes — based on confirmed word count and rate.
Done
Stage 11Project archived with full audit trail — every status change timestamped with the triggering function.
Built for scale, not for demos.
Every feature below runs in production at Liontalk, handling real AWS and enterprise client workflows daily.
GPT-4o Classification
Reads and classifies every incoming client email. Extracts project metadata, document type, and word counts from attachments — no manual data entry.
Intelligent Assignment
Sends translator offers in priority order, handles accept/decline via inbox monitoring, and auto-advances on no response. Fully hands-off.
Human-in-the-Loop
Every client-facing email is held for ops manager review before sending. Full control without the coordination overhead.
Real-Time Dashboard
Live view of all active jobs — status, assigned translator, hours in current stage, and stuck job alerts. No more status-check emails.
Concurrency Control
ETag-based optimistic locking prevents duplicate emails on concurrent function invocations. Zero duplicate sends, zero lost replies.
Event-Driven Architecture
Each pipeline stage triggers automatically on state change. Scales to any volume without adding headcount.
Built for agencies
that want to scale.
- Translation agencies managing 10+ projects per week
- Operations teams spending hours daily on email coordination
- Agencies looking to grow without proportionally growing headcount
- Agencies serving enterprise clients with tight deadlines
See it in action.
Book a 30-minute demo. We'll walk you through a live pipeline run, show you the dashboard, and answer every question about what it takes to integrate into your existing workflow.