For AMD Transportation, an AI booking agent reads job emails from shipping agents, turns their shorthand into bookings and tells the office on WhatsApp. Jobs are assigned to drivers, who acknowledge, update their status and capture signatures on the phone. Marine-Group Transport (MGT) runs its jobs, drivers, clients and invoices in its own version. Prestigze Asia's HR system handles rosters, shift reminders, site check-in with GPS, lateness and pay with overtime.
| Clients | AMD Transportation and Marine-Group Transport (crew transport for the marine industry), and Prestigze Asia (manpower supply), Singapore |
|---|---|
| What we built | A separate system for each company: bookings, driver jobs, vouchers and reports for transport; rosters, attendance and pay for manpower |
| Automated | Booking emails turned into jobs, driver assignment, WhatsApp job links and reminders, delay notices, shift reminders, no-check-in alerts, weekly schedules, reports |
| AI used | Google Gemini: AMD's booking agent, a recruitment screening assistant and a question-and-answer tool over roster and pay data |
What the systems had to handle
- Booking requests from shipping agents arriving as short, coded emails that had to be read and keyed in by hand.
- Drivers assigned fairly, told about each job and kept on time for vessel schedules.
- Staff rosters across client sites, with changes communicated quickly and acknowledged.
- Attendance, lateness and overtime worked out correctly for each person's pay.
How a crew transport job flows (AMD)
- A shipping agent emails a booking. The system checks the job mailbox every few minutes.
- The AI reads the shorthand. It turns the email into structured jobs: vessel, pickup, drop-off and timing. Follow-up emails in the same thread update or cancel the job.
- The office is told. An admin gets a WhatsApp summary. Emails the AI cannot read go to a review folder for staff.
- A driver is assigned. Automatically in turn with a rest gap between jobs, or by the admin.
- The driver acknowledges. They get the job on WhatsApp and are reminded if they have not acknowledged it within 10 minutes.
- The job is tracked. The driver moves it from assigned to on route, arrived and completed, and captures a signature. Late arrivals trigger a delay notice.
- Vouchers and reports. Each job's voucher is numbered and exported weekly, and a monthly driver performance report is produced.
How a roster flows (Prestigze Asia)
- Rosters are built. Weekly schedules are generated from shift groups, or uploaded from Excel.
- Changes are confirmed. Staff whose shift changes within 72 hours get a WhatsApp and must acknowledge it; admins are alerted if nobody does within 2 hours.
- Staff are reminded. WhatsApp reminders go out at shift start and end, and each person gets a weekly link to their shifts.
- Staff check in on site. By scanning a site QR code with their phone and GPS, without logging in.
- Lateness is flagged. Late arrivals and early departures are marked after a 15-minute grace period, and missing check-ins trigger an alert.
- Pay is calculated. Hours, monthly or hourly rates and overtime at 1.5 times, with overtime only counted when HR enters it.
What runs automatically
| When this happens | The system does this |
|---|---|
| A booking email arrives (AMD) | Reads it with AI, creates or updates the jobs and sends the office a WhatsApp summary |
| A driver messages a booking or photo (AMD) | Works out whether it is a new booking, an update or a status check, and holds it for confirmation |
| Jobs need drivers (AMD) | Assigns them in turn, with a rest gap for each driver |
| A job is not acknowledged in 10 minutes | Reminds the driver on WhatsApp |
| Every night (AMD) | Sends each driver a link to their jobs |
| A driver is late or still on site too long | Sends a delay notice |
| Upcoming jobs (MGT) | Sends WhatsApp notices to the people involved |
| A shift starts or ends (Prestigze) | Sends WhatsApp reminders, and an alert if nobody has checked in |
| A roster changes at short notice | Asks affected staff to acknowledge and alerts admins if they do not |
| Each week or month | Exports vouchers and builds driver performance reports |
What stays with the team
- Reviewing booking emails the AI could not read, and retrying or moving them.
- Letting an already-assigned driver know about a change.
- Assigning drivers by hand when needed, and approving final vouchers.
- Making the final decision on job candidates after AI screening.
- Entering overtime: only overtime HR types in is paid.
- Reviewing attendance flags before they affect pay.
AI that hands over to people
The AI does the reading and sorting: it turns coded emails and chat messages into bookings, screens job applicants against basic requirements, and answers questions about roster and pay data. Anything unclear goes to a person, and the question-and-answer tool is only given the rows it needs, never phone numbers.
Phone numbers are stored encrypted, every user has a role that controls what they can see, and automated jobs and incoming WhatsApp messages are verified before the system acts on them.
What we built
- AI booking agent for job emails
- AI handling of WhatsApp bookings and photos
- Automatic driver assignment
- Driver job links, acknowledgements and reminders
- Live job status and signature capture
- Delay notices
- Vouchers, weekly exports and driver performance reports
- Invoicing (MGT)
- Employee profiles and roles
- Shift groups and weekly schedules
- Excel roster upload with change acknowledgement
- QR and GPS site check-in
- Late and early-leave tracking
- Pay and overtime calculation
- AI recruitment screening
- Ask questions about roster and pay data
Could this work for your business?
The same booking, assignment and tracking setup fits any business that sends drivers, technicians or crews to jobs, and the roster and attendance side fits manpower agencies and multi-site teams. Implementation starts from S$3,000, plus a monthly fee based on usage, and a typical system is ready for testing in 2 to 4 weeks. Not sure where to start? Read what to automate first.
Related: all case studies, AI automation in Singapore, workflow automation in Singapore, how a project runs.
Still keying in bookings from emails?
Tell us how a normal day runs today. We will show you which steps can run by themselves.
Message us on WhatsApp