High-volume document intake
Handle 10,000 proof-of-delivery pages per month without forcing the office team to manually open, sort, and prepare every file.
Customer Story / Logistics
HANDWRITTEN + STAMPED POD AUTOMATION
From handwritten driver documents to verified client-ready PODs
DONE Deliveries is a European logistics company specializing in on-demand road transport solutions - from vans to rigid trucks and trailers - combining technology with hands-on operational support.
DONE Deliveries turned a labour-heavy proof-of-delivery check into a controlled automation workflow. Multibot reads handwritten and stamped driver documents, validates them against delivery rules and TMS records in an Exception-Led POD Verification for Driver Documents.
- Messy documents in, verified proof out. Photos, handwriting, stamps, and signatures become trusted proof-of-delivery records.
- Complex criteria checked automatically. Evidence, document type, required fields, and TMS matches are validated consistently.
- Humans review only the exceptions. Low-confidence, missing-stamp, unreadable, or mismatched cases are isolated for admin review.
PROCESSED SO FAR50,029 pages
Processed through the automated workflow to date.
SUCCESSFULLY PROCESSED98.5%
Processed successfully end to end.
AUTOMATED WORKFLOW100%
Of documents run through the automated workflow.
Included in every solution
Every Multibot solution comes with its own administration portal
The agents do the routine work in the background. The portal is where your team sees every item they handled, reviews the exceptions, approves changes and adjusts settings, with a full history of who did what.
- Live status of every item, from intake to completion
- Exceptions queued for your team's review and approval
- A full history of agent and team actions
- Users, rules and settings managed by your team
The bottleneck was handwritten, stamped validation, not just OCR
Each page had to be checked against business rules before it could be trusted
The challenge for DONE Deliveries was not simply reading text from a page. Drivers were sending phone photos of handwritten and stamped delivery documents in different formats, quality levels, and orientations. Some were CMRs, while others were supporting documents that still needed to be recognized, filtered, or routed correctly.
Before automation, the team had to manually interpret messy paperwork, confirm stamps and signatures, identify document types, match delivery data against the TMS, and decide whether each record satisfied complex acceptance criteria. That created a repetitive and error-prone admin bottleneck around a process that needed to run quickly, accurately, and at monthly scale.
- Driver-submitted photos arrived with inconsistent quality, rotation, handwriting, and stamps.
- The system needed to distinguish CMRs from supporting files before extracting handwritten delivery fields.
- Stamps, signatures, dates, references, parties, and TMS records had to satisfy complex validation criteria.
- Missing stamps, unreadable handwriting, low confidence, and mismatches needed to surface as admin exceptions.
- Final proof-of-delivery packets had to be prepared consistently for client handoff.
Automation Flow
How the workflow now runs
Multibot turns the hard handwritten-and-stamped validation work into a governed workflow. The detailed view below shows how driver photo intake moves through correction, classification, extraction, criteria checks, TMS matching, exceptions, and final proof-packet creation.
- 01
Driver submits delivery documents
Drivers send proof-of-delivery documents as photos, including CMRs and related supporting files.
- 02
Email and API intake
Documents are received through the existing intake channels and routed into Multibot automatically.
- 03
Image correction and preparation
Multibot prepares each image for processing, including rotating misaligned photos and improving the readability of handwritten and stamped documents.
- 04
Handwriting, stamp, and document extraction
The system identifies whether the file is a CMR or another document type, then extracts relevant handwritten fields, stamps, signatures, and delivery evidence from valid documents.
- 05
Complex validation, TMS matching, and admin review
Extracted data is checked against required criteria, stamp and signature evidence, and transport-management records. Exceptions, uncertain records, or non-CMR documents are surfaced for admin review.
- 06
Final proof packet creation
Valid matched documents are prepared for the final proof-of-delivery packet, ready for client handoff.
Capability Stack
What Multibot automated for DONE Deliveries
This was not a single-model OCR demo or a simple data-entry script. Multibot delivered a production document workflow for handwritten and stamped proof-of-delivery paperwork, covering intake, image preparation, classification, extraction, complex validation, matching, exception handling, and administration visibility.
Image correction before processing
Automatically prepare driver-submitted photos by rotating misaligned images and improving the readability of handwritten and stamped documents before extraction.
CMR and non-CMR classification
Identify whether each submitted document is a CMR or another document type, so the right files move forward and exceptions are routed correctly.
TMS matching and validation
Validate stamps, signatures, references, dates, parties, and extracted delivery data against transport records before a document is accepted into the final workflow.
Exception isolation
Surface missing stamps, unclear handwriting, low-confidence records, unmatched documents, missing data, and non-CMR submissions for admin attention.
Admin workflow visibility
Give DONE Deliveries an administration workspace where the team can follow processing status, review exceptions, and keep control of the full proof-of-delivery process.
Portal Intelligence
An administration workspace on top of the automation
DONE Deliveries did not just need background processing. The team needed a way to supervise the automation, understand where documents were in the workflow, and intervene only when required.
Multibot provided an administration layer where users can follow document intake, processing status, classification results, validation outcomes, TMS matching, and exceptions. The team stays fully in control, with visibility across every record - successful or failed - while the system highlights the exceptions that need human attention.
- Track documents from intake through final proof-packet preparation.
- See whether a file was classified as a CMR or another document type.
- Review image-processing, extraction, and matching outcomes.
- Identify unmatched records, low-confidence cases, and manual-review exceptions.
- Keep operational control without returning to inbox-based document checking.
Client testimonial
“We’d happily recommend Multibot and are already exploring additional work with them.”
Outcomes
What DONE Deliveries got back
DONE Deliveries replaced repetitive manual document checking with a controlled automation workflow that processes high monthly document volume, corrects driver-submitted images, reads handwritten and stamped CMR evidence, validates records against complex criteria and TMS data, and gives admins visibility into the full process.
Processing
98.5% successfully processed
Documents are processed successfully end to end through the automated proof-of-delivery workflow.
Scale
50,029 pages processed so far
The workflow supports live proof-of-delivery volume without requiring the admin team to scale at the same pace.
Admin work
100% of documents run through the automated workflow
Multibot automates the proof-of-delivery process across image rotation, document identification, handwritten CMR extraction, stamp and signature validation, TMS validation, and packet preparation.
Control
Visible administration workflow
The team can monitor document status, review exceptions, and follow each record through the proof-of-delivery process.
Next Steps
Multibot can map your intake channels, document types, image-processing requirements, matching rules, exception thresholds, and admin workflow into one operating layer for logistics document automation.