The AIIM Blog - Overcoming Information Chaos

Are You Leveraging Machine Learning Capabilities?

Written by Kevin Craine | Jan 30, 2019 3:05:00 PM

There‘s a lot of excitement about Artificial Intelligence and business automation these days, and for good reason. Developments in AI — and its sidekicks “Deep Learning” and “Machine Learning” — bring the promise of transforming work as we know it. Those transformed work processes will operate in a completely different way: fully automated and autonomous, with smart machines doing the work. The vision is to free humans from performing mundane and repetitive business tasks and assist them with better access to better information to better serve customers and the business.

As a result, it seems like just about every technology product now has the artificial intelligence “label” attached to it. For C-suite executives and technologists today, the challenge is to move beyond the hype to use machine learning and automation in ways that make a real difference in the performance of the organization.

Industry Watch Report on ML

A new industry watch report from AIIM titled “Leveraging Deep Learning and Machine Learning Capabilities” brings us some valuable perspectives on ways to move forward. A survey of 195 organizations was taken using a web-based tool in late November 2018. The core areas of responsibility for survey participants were primarily in information technology, as well as from line of business areas, compliance, security and records management.

The report is packed with findings and statistics that should prove valuable to both consumers and suppliers of enterprise information management technology.

We asked our survey participants four key questions:

  1. Where do organizations currently stand with regards to their Machine Learning and Deep Learning initiatives? Is the interest real or hype?
  2. What kinds of processes will be the initial target for Machine Learning capabilities?
  3. What do organizations see as the primary drivers for a Machine Learning initiative?
  4. What spending plans do organizations have for Machine Learning?

Research Results

We can’t cover all the results in this blog, but let’s take a quick look at some of the key findings. Where do organizations currently stand with regards to their Machine Learning and Deep Learning initiatives? Organizations clearly see Machine Learning as a priority moving forward, with 81% of the organizations indicating that AI and ML technologies are key to their future technology and business planning. However, organizations are still in an early stage of adoption — 87% say they are still exploring their options or are very recent adopters. Companies are working to figure out where to implement AI/ML technologies, but many are still refining the “how.”

Targets for Process Improvement

Given the high degree of interest surrounding Machine Learning, it helps to think about some concrete use cases, so we asked: What kinds of processes will be the initial target for Machine Learning capabilities? The results fell into four categories:

  1. Loss prevention through analysis of usage patterns (40%)
  2. Text analysis for better content classification and categorization – metadata assignment (35%)
  3. Better automatic understanding of the context of a document (33%)
  4. Automatic identification and indexing of documents (33%)

The concept of a “citizen developer” is seen by the majority of our respondents as important to their process improvement plans and over 50% of organizations feel automation of compliance and governance is “highly important” or “a deal changer.”

Drivers for Innovation

As with any disruptive technology, the key to finding success is using the capabilities in ways that make a difference, so we asked: What do organizations see as the primary drivers for a Machine Learning initiative?

For 79% of organizations, the ability to turn unstructured information like documents, images, audio and video files into structured data is key, but 87% of organizations find this task challenging. There is a huge backlog of “undigested” content — content that is not currently addressable by Machine Learning engines – that must be converted from unstructured information into structured data. Simply finding data and getting data “in shape” is a significant obstacle.

Spending Plans for 2019

And finally, what spending plans do organizations have for Machine Learning? There was a lot of feedback for this question, so I recommend that you download the full Industry Watch Report for more detail.

In the area of content services, 71% of our respondents indicated that they plan to use ML techniques in records management and preservation. In the area of process services, 70% are looking to improve business process management with Machine Learning. For analytic services, 69% are aiming to use the techniques for better content analytics and semantics. It is interesting also to note that 53% of those planning to spend “more” or “a lot more” on RPA technologies are planning on investing similarly on low-code application platforms.

Learn More

To learn more you can listen to John Mancini discuss this study at length on the AIIM podcast, AIIM On Air. In that episode, John and I explore the aspects of ML and Deep Learning, how organizations are planning to adopt the techniques, and what you can do to be prepared and move forward productively. And don’t forget to get your copy of the latest AIIM Industry Watch Report Leveraging Deep Learning and Machine Learning Capabilities.