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.