The artificial intelligence (AI) subsets of machine learning and deep learning are poised to transform the business landscape over the next few years, according to a new study conducted by 451 Research and commissioned by Iron Mountain. They will do so by automating repetitive tasks, identifying patterns in massive stores of previously untapped unstructured data and turbocharging knowledge worker efficiency.
These technologies have revolutionary potential across functions and industries, the researchers assert. For example, human resources departments deal with enormous amounts of semi-structured and unstructured content in documents like resumes, performance reviews and interview notes. This type of software searches for patterns that could also identify people with complementary talents who would be ideal members of “dream teams,” sorting through applications to find candidates whose skills and personalities best fit the organization’s needs and culture.
So, what exactly is machine learning, and how can organizations use this technology to their benefit?
What Is Machine Learning?
Machine learning is a subset of AI that uses algorithms to scour information and identify patterns and correlations that they aren’t purposefully programmed to find. These algorithms work with data sets that are too large for humans to process, and their accuracy and performance improves over time. For example, machine learning software can pore over millions of medical records to identify possible adverse drug interactions that might elude scientists. It can then identify new interactions based upon the success of past results.
Machine learning algorithms can also create structure where none exists. For instance, data about individual customers can be aggregated and analyzed to reveal behaviors and preferences that customers don’t explicitly state. The software scours customer comments to identify keywords and patterns that indicate positive or negative sentiment, or scrutinize legal documents for terms that may put a party at risk.
Many organizations shy away from committing humans to certain tasks for fear of violating regulations or simply because the problem set is too large. Machine learning algorithms bring both anonymity and scalability to the problem, enabling companies to improve customer experience quickly and at scale.
A related discipline called deep learning mimics the thought processes of the human brain to identify patterns. Deep learning-based image recognition systems are already better at recognizing faces and items than humans.
Benefits of Machine Learning
One of the most exciting breakthroughs of machine learning is its ability to derive intelligence from unstructured data. Until recently, data processing systems have been limited to working with well-defined and clearly labeled data types, such as email addresses and SKU numbers. But most of the information that organizations have is in unstructured vessels like emails, text documents and video files.
The biggest benefits early adopters envision are in the areas of customer experience, competitive advantage and increased sales. A significant minority of respondents also believe the technology will enable them to respond faster to opportunities and threats, reduce error rates and limit risk exposure.
Only about one-quarter of the more than 1,000 respondents to the survey cited cost savings as a benefit. Rather than cutting jobs, they believe AI will amplify the capabilities of the workforce to enable faster and better decisions.
By reinventing customer experience and driving employee productivity to new heights, researchers assert that organizations will gain a significant competitive advantage. But only if they successfully unlock the power of data that’s buried in unstructured sources, like email messages and text documents. Organizations that unlock the value of unstructured data and apply it proactively can gain rapid competitive advantage without running afoul of regulations, the report asserts.
Ready to Roll
Organizations are eager to put these new capabilities to the test. More than two-thirds of those surveyed are either using machine learning or have it on their agenda.
Of those with hands-on experience, expectations are high — 58% expect machine learning to have at least a moderate or significant impact on their organization, compared to just 24% who believe the impact to be modest. And 69% expect that impact to be positive, compared to just 5% who are pessimistic.
Although these technologies are still in the early stages of adoption, organizations are reporting high levels of satisfaction with initial applications — an overwhelming 92% reported positive opinions about their projects’ success.
The Future Is Now
Organizations expect to dramatically increase their use of semi-structured and unstructured data over the next two years, the survey revealed. However, that type of data is often the least well-managed, whether for reasons of complexity, regulation or lack of resources, researchers said.
The race will go to the swiftest, 451 Researchers say. “Machine learning and automation will be fundamental to identifying, sorting, classifying and governing information at scale,” they write. “Unstructured and semi-structured data is a relatively untapped resource, but that won’t stay the case for long.”