About the significance of machine cognition


About the significance of machine cognition

January 2017

AI for Supply Chain - not without RPA!

ERP and Supply Chain management software suppliers are finding more uses for AI to augment the intelligence their systems generates for users. This isn't surprising. After all, even a relatively complex compliance process such as KYC (Know Your Customer) is often mandatory in some form for every debitor (depending on the industry, even creditors need to be checked) added to the database.

And while the tedious information gathering required for KYC has long been delegated to automation scripts or software robots, interpreting the data can only be automated to a certain point. Depending on the industry, the Compliance Officer is "standing with one foot in jail" - simple statistical analysis might not fit the bill here (would you risk several years behind bars on someone else's algorithm?).

Bring in AI to sift through the data and recognize patterns that compliance personnel is alerted to. While completely autonomous compliance processes are still a vision for the future, a lot of the second level grunt work can be done at light speed, to leave the real decisions to humans trained in law.

Where would those AI systems be, however, without the diligent hard work, 24/7, of software robots gleaning the base learning data from possibly hundreds of different databases, portals and systems, either freely available or behind paywalls?

To derive deep learning from the real world, it isn't enough to dump a few databases into a neural network - the data needs to be complemented by external sources and has to be updated continuously. Only RPA is able to collect the necessary depth and breadth of external data at a reasonable cost!

On RPA Strategy for large organisations

Implementing an RPA strategy isn't like a roll-out of the newest Office suite. RPA needs to be driven from C-Level management if it is to succeed. Unlike SMB's, where RPA may be used for a limited number of processes, a corporate implementation of the technology should be more highly structured.

This includes dedicating resources to the topic and setting up a center of excellence for RPA in the organization. It is imperative that an investment in RPA technology and resources be usable across all departments, from Marketing to Finance. RPA requires a consultative approach, in order to seek out as many processes that will benefit from automation as possible, but also to evaluate the economic aspects of automation.

Technical RPA infrastructure is - in most cases - either relatively simple to implement on-site or available as cloud-based services. The cost to RPA lies in selecting and analyzing process, in training personnel to set up robots and - often overlooked - in robot maintenance.

This Information Week article does a good job of covering the strategic aspects of RPA.