Machine Learning and Artificial Intelligence in Digital Document Management

The combination of Lemmana machine learning and Document Logistix’ Document Manager enables organisations to undertake digital transformation today and improve business performance as they use and teach the centralised system to manage documents quickly and accurately.

Excessive admin time and costly errors

Any time and motion study will reveal that the time and costs associated with capturing information received in an organisation can be staggering. Historically – and even today – the capture process involved opening envelopes, reading content, distributing the post and assigning tasks.

Even with electronic systems misfiling is a regular event at the point of capture, and missing or out-of-date documents frequently blight workflows.

The costs associated just with getting a business process started can be very high in terms of staff wages and time wasted looking for the right version of a document.

It’s generally thought that knowledge workers spend 20% of their time looking for documents

 

Expressed another way, it could be that some of your staff are only ‘working’ a three-and-a-half day week!

Document Manager: Making work flow

Document Logistix’ Document Manager software enables organisations to automate workflows from information capture, through pre-defined checkpoints, approvals and threshold alerts, to compliant retention and eventual data destruction.

The digital process provides management with line of sight, which enables project owners to take timely action to prevent bottlenecks and delays.

One of the most reassuring aspects of a digital workflow is the real-time availability of business-critical reports, such as task completion, procurement, production, queries and complaints, cash flow, user activity and audit trails.

Lemmana ‘on board’: The best start to the workflow process

A digital mailroom speeds up the process of indexing and assigning documents to the correct location, whether that be a folder, a department, a job function, a person or global dissemination.

Documents can arrive in the virtual mailroom in any electronic format, PDF, Word, Excel, HTML, photograph image or video. Or the provenance of a document may be paper, in which case it may be scanned individually or as part of a rapid batch-scan directly to the mailroom, and subsequently routed and distributed according to the document’s type, or any other classification.

Document Manager now has Lemmana ‘on board’, which adds a new layer of intelligence to help Document Manager users to reach the highest possible levels of speed and accuracy in the capture and distribution of documents.

Lemmana automatically categorises documents based on their profile and characteristics, and assigns them accurately to the next stage in the workflow. 

A document can be classified based on its content or layout and the respective metadata captured such as customer number, invoice number, name, date or car registration. 

Invoices, complaints, insurance records, vehicle fleet documents, proofs of delivery … all appear instantly at their next pre-defined destination for storage or processing.

Train your documents to behave

Lemmana’s machine learning capabilities uses Artificial Intelligence to enable Document Manager users to make incremental improvements to the document management system and, by extension, their business operations. 

Document Manager users can train the system to look for specific document types, or to identify content traits such as logos, to ensure that documents are distributed and assigned accurately. See Functionality Overview.

Users require no technical knowledge, and they can progressively train the system to eliminate ambiguities and achieve the highest levels of recognition and assignment accuracy.

No tech assistance needed: Self-sufficiency

Traditional capture systems typically require technical consultants to set-up multiple rules and code routines to capture data from documents. Lemmana is very different; its functions make the user’s life easier by bringing together Computer Vision, Machine Learning, Natural Language Processing (NLP) and Artificial Intelligence (AI) approaches into one process. Users train the system through a simple web-based interface about what types of document are received and what data is required from them.

In the background, the system builds capture models based on these user samples, which are then applied to incoming documents for automated classification and data capture. The documents can then either be verified and/or stored or matched to your workflow processes.

Folders and documents: Size no longer matters

The number of documents associated with conducting business, in the front office and back office, has grown phenomenally and poses a huge logistical challenge.

A typical customer record comprises everything from credit check and NDA to purchase orders, invoices, queries and credit notes.

An HR file for an individual employee comprises dozens of documents including job offer, induction notes, contract, identification, holiday requests, payment details, emergency contacts and more.

However, with Lemmama and Document Manager, the content within each document type can be stored accurately and retrieved instantly in a compliant format. You have the ability to search the system and retrieve the required content contained on pages 4, 18 and 43 of a 50-page document.

Compliance: Not all documents and data are equal

The advent of GDPR in 2018 heightened the already strict legislation surrounding the storage, retention and sharing of personal data.

What were once accepted practices suddenly became fraught with all sorts of compliance pitfalls. Take the seemingly simple matter of storing a passport or driver’s licence as part of your employee on-boarding process. What data must be stored? What information should be visible internally, and what data should not be made public, or shared as part of a subject access request?

Lemmana makes it possible to redact (digitally mask) sensitive data automatically in a fraction of the time it would take a human to achieve, and with far more accuracy.

Personal data: Don’t be haunted by the past

Redaction can be applied to the appropriate content fields or selected metadata within any document uploaded to the system or when it is viewed or retrieved later. The Compliance Officer and the Board can have confidence that they have built compliance best practice into their operations.

In the past five years, progressive companies have implemented digital strategies to replace physical archives, storage facilities, warehouses and filing cabinets. Document Logistix has helped many organisations with digital transformation: one project involved back-scanning 15 million HR records as part of a blue chip’s People and Culture Department’s drive to cope with the demands of GDPR. 

With Lemmana ‘on board’, documents can also be batch-scanned rapidly to a central hub with sensitive data automatically redacted.

The results of digitisation are positive across the board: organisations save money, reduce errors, increase productivity, and free staff members from mind-numbing admin tasks to work on higher value projects. 

Digital workflows also facilitate remote and home working, which enables companies to become more agile and reduce office overheads.

Overcome retention headaches

Document Manager enables you to simplify the complexities surrounding information retention regulations.

There are standard retention periods for certain types of document such as accounting records, which must be retained for seven years. Documents can be time-stamped to be purged from the system at the mandated time.

However, some records require part-retention and part-deletion. When an employee leaves an organisation, some information must be kept on file for varying lengths of time, while some must be deleted immediately, for example the employee’s emergency contact details.

To manage these retention/deletion periods manually is both difficult and time-consuming. Digital document management enables you to automate the appropriate actions based on triggers such as employment end-date.

Off the shelf document recognition to jump start the system

We can help you to train your digital document management system to recognise the documents that flow through your business on a daily basis, so that you achieve a streamlined, error-free workflow. 

See Document Manager Machine Learning Functionality Overview.

If you deal with a lot of standard document types, you can jump start the system using pre-built recognition formulas.

Watch our Artificial Intelligence User Group video