AI-Powered Invoice Extraction for Travel and Mobility Companies

Relevance Digital has delivered invoice-extraction solutions for a ferry-ticket booking company and a mobility leasing company in separate implementations. The solutions use Document AI and Gemini, provide access through an application programming interface (API) and include an interface for human review.

They apply AI to a defined task: extracting information from invoices while keeping people involved in reviewing the output.
 

Support invoice extraction and review

The objective was to extract invoice information in a form that other software could access, with a dedicated interface for people to review the results.
This brings document processing and software integration together. Extraction handles the information within the invoice, the API makes the output accessible and the review interface supports human oversight.
 

Invoice extraction using Document AI and Gemini
The solutions for the two companies use Document AI and Gemini to extract information from invoices.
The scope is document extraction. It does not imply that the applications approve payments, make accounting decisions or autonomously complete an entire invoice-processing workflow.

Access through an API
The extraction capability is served through an API, allowing other software to interact with it.
This provides a way to incorporate document extraction into a wider application or workflow. The specific downstream systems and connections are not detailed in the available project information.

A human review interface
A dedicated interface allows people to review the extracted information.
This keeps review within the solution rather than treating AI output as a final, verified result. The interface provides a point of human oversight before the information is used further.
 

From document extraction to review

The implementations combine three capabilities: AI-based invoice extraction, API access and human review.

Document AI and Gemini support extraction. The API provides a software connection to that capability. The review interface enables people to inspect the output.

The available information does not specify how invoices are routed for review, which validation rules apply or how corrections are handled. Those details are needed to assess the complete quality-control process.
 

AI as part of a document-processing workflow.

The work for these travel and mobility companies demonstrates how Relevance Digital combines AI services with application development and human oversight.

These are separate implementations. Each should be assessed against its own document types, integration requirements and review process.

Measured results for extraction accuracy, processing time and manual effort are not yet available in the supplied project information.