With the emergence of the Digital Age, nearly every business executive has felt the ground shift beneath his or her feet. Because the business landscape continues to change, steadying one’s self is not easy. Consultants are telling companies to transform into digital enterprises or risk being tossed into history’s dustbin. Key to making this transformation is embracing and leveraging cognitive technologies (i.e., artificial intelligence (AI) platforms). Machine learning is a central capability of all cognitive technologies. Kostiantyn Didur, a Senior Marketing Manager at N-iX, bluntly states, “Companies can’t turn a blind eye to machine learning anymore, as it is so powerful at certain tasks.”[1]

Getting started

Didur wisely cautions companies to do their homework before jumping into any machine learning project. He recommends asking and determining the answers to four questions. They are:

1. Do you really need machine learning? For many companies, the answer will be yes. Barbara Call (@BarbararRCall1) asserts machine learning will have the greatest impact in the areas of “security, data-driven decision making, and product development.”[2] Didur warns, however, “Machine learning projects are costly and take up much time. … Not all companies need machine learning in the first place. … Solid data engineering alone can be enough to find invaluable business insights for companies across numerous industries.”

2. What type of machine learning do you need? Cynthia Harvey explains, “All machine learning systems are AI systems, but not all AI systems have machine learning capabilities.”[3] With that bit of trivia out of the way, Harvey goes on to note there are four types of machine learning. They are:

• Supervised Learning: “Supervised learning requires a programmer or teacher who offers of examples of which inputs line up with which outputs.”
• Unsupervised Learning: “Unsupervised learning requires the system to develop its own conclusions from a given data set.”
• Semi-supervised Learning: “Semi-supervised learning, as you probably guessed, is a combination of supervised and unsupervised learning.”
• Reinforcement Learning: “Reinforcement learning involves a system receiving feedback analogous to punishments and rewards.”

Didur asserts, “Nearly 90 percent of current machine learning development projects deal with supervised learning.”

3. Are you ready for extensive data engineering? Machine learning and data have a symbiotic relationship. Without large amounts of data, machines can’t learn; and, without machine learning, insights can remain trapped inside data. Didur notes, “You need to extract data from fragmented sources, transform it into usable datasets, and load it to the AI engine. The bad news is these tasks often cannot be automated. Different sources have unique types of data, so adjusting them requires a lot of manual work.” Fortunately, cognitive computing platforms can work with both structured and unstructured data.

4. Do you need custom development, API software, or startup acquisition? Some analysts are concerned about the largest global companies acquiring all of the promising AI startups. Didur notes that acquisition is the costliest option for most companies. On the other hand, he notes, building a machine learning solution from scratch can be “the riskiest option, as only an estimated 10 percent of machine learning R&D projects succeed.” He recommends a middle path of working with AI vendors.

Once you successfully answer those questions, you can begin your work in earnest. According to Clint Boulton (@ClintBoulton), Gartner analysts recommend beginning with “small experiments in different business areas with particular AI technologies for learning purposes, not ROI.”[4] At Enterra Solutions®, we recommend a “crawl, walk, run” approach so solutions can be tweaked and tuned before going full scale.

Achieving success

Despite the obvious benefits of machine learning and other cognitive technologies, Ramon Chen (@RamonChen), Chief Product and Marketing Officer at Reltio, cautions a lot of hype remains. To ensure you get the most from machine learning projects and don’t fall victim to the hype, he offers 3 critical ingredients for your consideration. They are:

1. Create a reliable data foundation. As noted above, machine learning and data have a symbiotic relationship. Chen notes, “Most companies are NOT ready for any form of AI, ML, or Cognitive Computing to help their business user, because their data is such poor shape to even attempt such an endeavor.” A good vendor can help you work through your data challenges.

2. Bring analytics and machine learning to the data. Call notes, “AI and machine learning technology can handle volumes of data that the human mind can’t fathom — and in this era of big data, data analytics, and business intelligence, it’s clear that AI and machine learning have a role to play.” Chen warns, “Bolting on AI or ML into legacy master data management (MDM) systems, or using such MDM tools to feed downstream disparate ML tools is putting lipstick on hosted managed services disguised as cloud. Reliable data, relevant insights and recommended actions via machine learning needs to be seamlessly combined into one, single multi-tenant cloud platform, architected from the ground up, for both analytical intelligence and operational execution, through data-driven applications. Successful execution requires a closed-loop of all data, insights and actions, to ensure accurate metrication for continuously improved outcomes.”

3. Don’t go all in on one vendor. Chen argues, “In a rush to market that ‘our tools do it too,’ large vendors will unfortunately, over promise, and under deliver. It’s not their fault, as they must respond to the market, but many face an unenviable task of achieving ingredient #2 above, let alone attempting to now also execute on a plan to deliver their own AI technologies.” The point is, you must find vendors who are adaptable and responsive to your unique challenges.

Get the process right and good things will happen. The important point to note is that gaining insights in the Digital Age is a “process” not a solution per se. Didur observes, “Data science and machine learning often produce unexpected results and give invaluable insights. This technology is here to stay, and it’s going to evolve at an extremely fast pace.” Your company needs to embrace an adaptable cognitive system that can change with the times.

Summary

Call concludes, “Technology that’s equipped with AI or machine learning can be a powerful tool for the kinds of tasks humans just can’t master. Machines can recognize sequences in large volumes of data, identify behavior that doesn’t fit normal use patterns, or apply speech recognition to unusual applications. But they don’t have the general intelligence that humans do — which is why AI and machine learning will likely augment human skills, not replace them outright.” Machine learning is poised to become one of the most important tools in many companies’ business kit.

Footnotes
[1] Kostiantyn Didur, “4 Questions to Ask Before You Start a Machine Learning Project,” Business.com, 9 July 2018.
[2] Barbara Call, “AI and Machine Learning – The Potential to Challenge and Transform,” CIO, 14 December 2017.
[3] Cynthia Harvey, “What is Machine Learning?” Datamation, 3 January 2018.
[4] Clint Boulton, “10 tips for getting started with machine learning,” CIO, 12 September 2017.
[5] Ramon Chen, “Three Critical Ingredients for AI, Machine Learning & Cognitive Computing Success,” Reltio Blog, May 2017.