It’s time to move beyond simple streaming analytic applications to truly cognitive applications that incorporate the latest in machine learning and deep learning.
What is continuous intelligence?
In animals, the nervous system continually receives input from various sensors – audio, visual, temperature, and more. Over time, animals learn behavior and how to react, such as swatting a mosquito that’s landed on your arm or awaking from a deep slumber at an infant’s cry. More than 99% of this continuous sensory input is discarded as irrelevant. Some information is retained in short-term memory, like what you ate last night. Other things like an old friend’s face make it to long-term memory, recalled nearly instantly when seen.
Today’s leading organizations are taking a page from nature. They have deployed continuously intelligent systems that rely on platforms, architectures, and software using real-time and historical data to enable critical decisions and actions to be taken in milliseconds to minutes. Continuous intelligence (CI) means learning from new and historical data using artificial intelligence to gain insights and act appropriately. CI goes beyond streaming analytics, which generally only apply a few filters, transformations, or aggregations. Like the nervous system, CI propels people to act immediately if necessary, or later as appropriate.
Why continuous intelligence matters
People have come to expect instantaneous answers due to near-universal access to smartphones and the internet. People long for having their every need anticipated and met. Those companies that learn to meet and exceed expectations in a timely fashion will not only grow their customer base but gain more value from each customer.
In a recent IBM Institute for Business Value Research Insight, six strategies of top leaders were highlighted to perpetuate success. One strategy was to curate data that “thinks” and “acts.” That is, turn your data into immediate and actionable insights using technologies like AI and automation.
“Tide and time wait for no man,” said Geoffrey Chaucer more than 600 years ago. What was true for people then is just as important today for businesses. Taking longer to discover and resolve issues increases cost and waste in a business. Whether it’s equipment malfunction leading to waste on the shop floor, or klunky software interfaces driving away consumers, the result is decreased profits. Immediate action to detect and resolve anomalies allows just the opposite – improved profits through cost savings and revenue growth.
Continuous intelligence in Action
Sophisticated, continuously intelligent applications use the widest range of artificial intelligence, including both machine learning and deep learning to deliver unsurpassed value. Here are a few examples:
One IBM customer has applications in production with more than 1,700 AI models for customer care. They anticipate and divert call center calls to lower-cost support channels. They predict successful ads, leading to significant increases in web advertising revenue. Their continuous intelligence applications have also increased revenue through cross-selling.
Another company continuously monitors everything about their transportation operations, from crew and passengers aboard each plane to hours worked each day and month. When irregular operations cause flight delays, the company immediately determines if a crew would violate United States Federal Aviation laws on how long a crew can work each month or day. With insight based on passengers aboard, flight connections and past operations, it can decide if it’s more profitable to cancel the flight or call in the backup crew. The immediate action contributes to a 150% return on investment.
Tide and time don’t wait. Nor do competitors or customers. Early adopters have proved the capabilities to create continuous intelligence systems in mission-critical areas. It’s time to move beyond simple streaming analytic applications using simple transformations and enrichment to truly cognitive applications that incorporate the latest in machine learning and deep learning.