Tuesday, September 30, 2014

The driverless ECM

Humans have been convinced for years that the car industry was really advanced. We have been proud of "smart" cars that were able to read traffic sings, alert in short distances and tell you to buckle up.

That was until the presentation of driverless car in 2011 by Google, when we all saw the real automation in car industry. This is a really smart car, this is a real milestone for an industry, giving humans the possibility to move without manual tasks that lead to many errors and dangerous situations. Since 1886 when the Motorwagen was invented, it has taken the industry more than 130 years to reach this magic.

The ECM industry is evolving really fast, but we are still anchored in "smart" features that are really far from the real "driverless ECM”.

Current ECM solutions need you to drive them during all their life to go anywhere you want to go:
  • They need you to create, and sometimes code your document types
  • The same goes for metadata, types of metadata, schemas, layouts, etc.
  • They require you to do the whole hierarchy and structure of your documents organization
  • And all of this, you have to agree with your company’s team to create an effective organization of the brand new ECM solution :-)

Many of you surely know what I mean. The process to set up an ECM system in any organization is like driving the old Route 66 in a Chevrolet bel air. Nice but zero automation.

Many experts would mark the ECM market a mature one, but until we get the real automation, we can compare it to the mature industry of manually driven cars.

ECMs current industry lacks a few gadgets for the driverless challenge:

A high definition map of the document type, metadata and features related to the ECM world map.

An important step for this milestone will drive us through a deep process of “feature engineering” and to have the result of it as a “knowledge base” from which any of the existing ECMs would be able to take information and allow users to reuse and even propose them their document types, metadatas and overall configuration.

A good light radar to measure the distance (and other things) of documents, metadatas and features.

In ECM operations we are commonly asking “are these documents the same?”, “are they similar?”, “do they have the same value for some metadata?”, many of this questions are typical searches for any ECM user. The ECM industry, still lacks a standard in comparison algorithms, solutions and standards.

A very precise range finder laser to model the real world around a document.

This is what today we call “Document Analysis Systems”. There is a really incredible work done right here, but the only solution that seems really industrialized is OCR. For layout recognition, language detection, decoding and some other features, there are solutions, but none of them really reaches this milestone yet.

Some computation pieces to decide where to go and what to do with the documents.

ECM and BPM worlds have a permanent affair with some ups and downs. There are really complex workflows in the nature of ECM, not because of the workflows but the elements that interact, security issues, external and internal users, metadata updates, tasks and many side-effects that take when a document goes from one state to another.
The ECM industry really needs to create a “driverless experience” in this affair. Users deserve to count with existing workflows, ECMs that suggest a workflow to use, a route to take. Imagine your ECM telling “This is an invoice, do you want to send it to the accounting department?”

A remote farm of computers to do the complex tasks

Driverless car has been possible thanks to many of the now existing technologies and one of them is the ability to do distributed computing for complex tasks. If the driverless car is doing it now, the ECM world must do it now. Some of the “smart” features that we expect from a driverless ECM really require a hugh amount of computing, lots of image processing, lots of text processing, lots of machine learning that needs to be done outside. The ECM world, needs a standard to get rid of heavy tasks and focus on the driverless experience of the user.

Real automation is still very far from ECMs, how much time will it take us to see the driverless ECM?

Jose Luis de la Rosa
CEO at Athento