Currently set to Index
Currently set to Follow

Document processing thanks to Artificial Intelligence

Prepare any document for further processing with the usage of artificial intelligence

The natural result of transferring our reality to digital world is a rapid increase of data quantity. By using just the regular ERP or ECM software that requires employees supervising, it is impossible to effectively process the informational chaos. A handy solution in this case is a contextual software based on the artificial intelligence.

Module Artificial Intelligence will support you in any document preparation for further processing, ex. it will recognize its type, verify data correctness, export data to desired fields in the ECM or ERM system formula or will prevent duplicates and alarm about unnatural, dangerous operations.

Document processing thanks to Artificial Intelligence

Prepare any document for further processing with the usage of artificial intelligence

The natural result of transferring our reality to digital world is a rapid increase of data quantity. By using just the regular ERP or ECM software that requires employees supervising, it is impossible to effectively process the informational chaos. A handy solution in this case is a contextual software based on the artificial intelligence.

Module Artificial Intelligence will support you in any document preparation for further processing, ex. it will recognize its type, verify data correctness, export data to desired fields in the ECM or ERM system formula or will prevent duplicates and alarm about unnatural, dangerous operations.

How does document processing work in the Artificial Intelligence module?

With the usage of electronic document workflow the purchase invoice will automatically end in the accounting department , CV to HR department and offer to secretary.

Scan or photo of the invoice or other document

OCR recognizes characters, accessing the whole document

Data validation

Automatic classification or data extraction

ERP/ECM/Workflow
BPM
DMS
RPA
Contextual search
Other applications

Scan or photo of the invoice or other document

OCR recognizes characters, accessing the whole document

Data validation

Scan or photo of the invoice or other document

ERP/ECM/Workflow
BPM
DMS
RPA
Contextual search

During the manual document implementation or automatic mail downloading a OCR process occurs, what means that the data is converted into text and converted into PDF file. Software can detect document type and automatically send it to the correct department or person, or just suggest the right receiver with the percentage of correct recommendation possibility.

Further document processing steps can differ depending on service use scenario. Data recognized from the invoice can be automatically imported into form in the ERP system, received CV will be sorted according to the vacancy, removing all the duplicates and detecting key words to make selection easier.

The most important components of AI module for document processing

OCR

A service which adds text layer into graphic files (*.jpeg, *.png or unsearchable *.pdf). Any file from the scanner or document picture will be converted into searchable *.pdf file.

TEXT CLASSIFICATION

A service which recognizes document class based on the historical data. It allows for automation of routine and repeatable tasks, ex. recognizing the document type that was sent to firm mail and distribution to the right department.

DATA CAPTURE

A service which enables for data extraction from a document to the systems form. Algorithms check the correctness of returned data, so they can validate those information, ex. VAT, TINs, SSNs or bank account number. We can define document template or use already existing ones, which doesn’t require users work: data capture invoices, business cards, tables or “key-value” pairs.

ANOMALY DETECTION

A service which detects anomalies and non-typical situations. For example it can be used for invoice verification to check if the costs from the outsourcer doesn’t differ from expected values, or to monitor non-typical user activity in the system (uploading too many documents, or downloading too many files).

TEXT SIMILARITY DETECTION & ANALYSIS

A service, which enables intelligent text analysis, including the words (token) meaning and its place inside the sentence and the whole text. It also allows for recognition of similar texts, fragment duplications and even generation of key phrases and document summaries.

The most important components of AI module for document processing

OCR

A service which adds text layer into graphic files (*.jpeg, *.png or unsearchable *.pdf). Any file from the scanner or document picture will be converted into searchable *.pdf file.

TEXT CLASSIFICATION

A service which recognizes document class based on the historical data. It allows for automation of routine and repeatable tasks, ex. recognizing the document type that was sent to firm mail and distribution to the right department.

DATA CAPTURE

A service which enables for data extraction from a document to the systems form. Algorithms check the correctness of returned data, so they can validate those information, ex. VAT, TINs, SSNs or bank account number. We can define document template or use already existing ones, which doesn’t require users work: data capture invoices, business cards, tables or “key-value” pairs.

ANOMALY DETECTION

A service which detects anomalies and non-typical situations. For example it can be used for invoice verification to check if the costs from the outsourcer doesn’t differ from expected values, or to monitor non-typical user activity in the system (uploading too many documents, or downloading too many files).

TEXT SIMILARITY DETECTION & ANALYSIS

A service, which enables intelligent text analysis, including the words (token) meaning and its place inside the sentence and the whole text. It also allows for recognition of similar texts, fragment duplications and even generation of key phrases and document summaries.

OCR speed
OCR efficiency
TEXT CLASSIFICATION speed
TEXT CLASSIFICATION efficiency
DATA CAPTURE speed
DATA CAPTURE efficiency
OCR speed
OCR efficiency
TEXT CLASSIFICATION speed
TEXT CLASSIFICATION efficiency
DATA CAPTURE speed
DATA CAPTURE efficiency

Scenarios of AI module implementation

OCR on own server

Processing pictures into searchable pdf file slowly becomes a standard. However this process is often being distorted (OCR on the scanner side, other applications etc.). in our model the OCR service works on the server side, it is effective because of the container structure (docker) and allows every document to be in searchable format.

Contextual searching from OCR documents

You can create advanced workflows between singular employees. System enables form document management in different process phases: data hiding, edition blocking or enforcing fields filling.

Ready to use data capture models

Data capture models are ready to use without further training. The most often our clients use this model for trade documents (invoices), but we are working on new ones (business cards, CVs etc.).

Creating own data capture models

You can teach general models or create new ones from scratch. In the case of model teaching (ex. for invoices) we configure settings of existing fields or we add handling of new ones. Settings can be saved for particular contractor. You can also create handling of new document classes.

Document type recognition

The basic use of text classification is selecting a document type.  Based on the documents text, the system decides if it is dealing with the invoice, contract or script. Training of type classificator is individual for every environment.

Form fields recommendation

Other implementation of text classificator is predicting the value for choice fields. System analyses previous documents and can predict the correct value for a field like: documents owner, accounting records or cost centrum.

Scenarios of AI module implementation

OCR on own server

Processing pictures into searchable pdf file slowly becomes a standard. However this process is often being distorted (OCR on the scanner side, other applications etc.). in our model the OCR service works on the server side, it is effective because of the container structure (docker) and allows every document to be in searchable format.

Contextual searching from OCR documents

You can create advanced workflows between singular employees. System enables form document management in different process phases: data hiding, edition blocking or enforcing fields filling.

Ready to use data capture models

Data capture models are ready to use without further training. The most often our clients use this model for trade documents (invoices), but we are working on new ones (business cards, CVs etc.).

Creating own data capture models

You can teach general models or create new ones from scratch. In the case of model teaching (ex. for invoices) we configure settings of existing fields or we add handling of new ones. Settings can be saved for particular contractor. You can also create handling of new document classes.

Tworzenie własnych modeli przechwytywania danych

Document type recognition

The basic use of text classification is selecting a document type.  Based on the documents text, the system decides if it is dealing with the invoice, contract or script. Training of type classificator is individual for every environment.

Form fields recommendation

Other implementation of text classificator is predicting the value for choice fields. System analyses previous documents and can predict the correct value for a field like: documents owner, accounting records or cost centrum.

Other functionalities

Want to learn more about the AI module possibilities?





    Contact form is the best way to get in touch with us. Your message will reach one of our Sales Leaders, who will respond to your inquiry at lightning speed. You can also call us or choose to talk to us via social media: Facebook and Linkedin.

    Mateusz Illg

    Key Account Manager
    Menu