The AI And Machine Learning Nirvana: Reachable With The Right Approach Posted on : Jan 25 - 2019

There’s a perception out there that two evolving technologies, artificial intelligence (AI) and machine learning (ML), will provide businesses with greater efficiency, increased productivity and a beefier bottom line — almost overnight. And with that message, companies are ready to jump in.

According to a 2018 Gartner press release, global business value from AI was expected to reach $1.2 trillion in 2018, a 70% jump from 2017. By 2022, Gartner expected it to reach $3.9 trillion. The research firm predicted that AI will be “the most disruptive class of technologies" in the next decade.

As the CTO of a company that provides AI-driven end-user experience monitoring and cloud solutions, I don't believe there's any doubt that AI and ML are technologies that will bring a lot of opportunities for organizations to leverage data and enhance everything from sales to marketing to operations. The technology is attainable, and the potential is limitless. But the digital nirvana that AI and ML may promise won’t just come out of a box. Organizations will need to invest the research and time it will take to strategize their AI/ML investments, to identify the sources of their data and to prepare to shift gears as the dive into big data becomes more revealing.

Lower Barriers To Entry

The good news is that factors like the cloud and open-source availability have lowered the barriers to entry for companies looking to explore the benefits of an AI and ML play. Thanks to the cloud and the availability of pay-to-play cloud services, the days of hardware and software chasing each other’s upgrades — and the compatibility issues that came with that back-and-forth — have largely gone away. Open-source is another reason I see for the faster adoption of AI and ML. Python libraries like TensorFlow (developed by Google) allow companies to leverage 1,000 or more contributors’ work.

But before they head down that road, they need to map the journey. They need to ask themselves where the data will come from — and what they’ll do with it once they learn more about it. Once data is normalized, analyzed by a company's technology and becomes more granular, it’s sure to generate even more data. In the past, working with massive datasets often required dedicated compute farms. Today, with open-source solutions like MongoDB or Elasticsearch with Kibana, companies can try out different strategies before committing to a full-blown adoption. Visualizations provided by Kibana allow humans to see if a dataset passes the smell test. For example, the peak website traffic should generally follow the business hours for that region. A new launch of a product should drive up the transaction counts on the web and backend servers. More importantly, the transactions, as observed by application and network visibility tools, should line up. If they don't, it’s a sign that the tools are not properly synchronized, and visualizations can help companies "smell test" datasets like these. View More