Virtualization Technology News and Information
Article
RSS
Why "Low Risk" AI can be an entre to more advanced technology

 

Written by Daniel Fallmann, CEO of Mindbreeze

In 2018, Google CEO Sundar Pichai said, "AI is one of the most important things humanity is working on. It is more profound than electricity or fire." In the past few years, it has matured beyond the "what ifs" to a point where real, tangible results can be expected from the technology. Oddly enough, despite the widespread acceptance of AI's potential to drastically impact the way we do business, there is still hesitation to adopt the technology itself. A study by Gartner found that AI adoption has tripled in the last year alone, but only an estimated 37 percent of firms are now implementing AI in some form.

Although we all agree that AI is important, businesses across industries are still looking for ways to make it work for their business only to come up short in the process. Knowing where to start and how to profitably employ it is not always clear. For those who are just getting started, a high-budget, business-critical program may not be the best place to begin.

One of the biggest stumbling blocks around AI-adoption is the misconception that understanding human language is already a completely solved problem. The industry is already technologically advanced in semantic analysis, and interpreting a variety of languages is already possible. However, the accuracy of self-learning methods is strongly dependent on the underlying information that is not always clear-cut or easy to interpret. That's why low risk, non-business critical solutions that truly understand an organization's communication structure can serve as an AI proving ground.

Applied AI is not industry-specific, but rather department-specific. If one has a successful start with a business case, the acceptance of the users and the respective company is quickly given. For example, customer service has a range of use cases, including processing incoming mail, classifying information, and communicating existing best practices and existing knowledge.

To dive deeply into potential "low risk" applications, let's look at the aforementioned example: incoming mail/classification of information. How can companies improve the way they process incoming communications? There are many different input channels, such as the mailbox, the email system, or social media channels in any company. As a result, more companies are interested in finding ways to classify incoming communications. Automated incoming mail classification is made possible by natural language processing (NLP) - a lower risk AI solution. Unstructured communications in particular are becoming an increasingly common area. Only through methods like NLP can a content-based and form-free analysis become truly automatable for all input channels. Employees are no longer required to read digital communications independently and forward them to the right department.

There are many ways to find your own AI proving ground. A successful use case with "low risk" AI like natural language processing - technologies that have already begun to mature and prove their value - can serve as a non-business critical use case showing that the technology can help achieve measurable business goals.

##

About the Author

Daniel Fallmann founded Mindbreeze in 2005 at the age of 23 after finishing his studies in computer science. He has many years of experience in the computer and information technology sector. As Mindbreeze's CEO he is a living example of high quality and innovation standards. From the company's very beginning, Fallmann, together with his team, laid the foundation for the highly scalable and intelligent Mindbreeze InSpire appliance and cloud service. His passions for enterprise search and machine learning in a big data environment have fascinated not only the Mindbreeze employees, but also their customers.

Published Tuesday, June 04, 2019 7:25 AM by David Marshall
Filed under: , ,
Comments
There are no comments for this post.
To post a comment, you must be a registered user. Registration is free and easy! Sign up now!
Calendar
<June 2019>
SuMoTuWeThFrSa
2627282930311
2345678
9101112131415
16171819202122
23242526272829
30123456