Predictions 2020: What to Expect in AI, ML, RPA Sectors

At the end of each year, eWEEK likes to post observations from IT thought leaders about what they think we should all expect in the coming year—new products, innovative services, trends to look for and so on. As usual, we’ve published a few of these thus far, but we had some publishing-process snafus in December and early January, and many of the remaining articles didn’t get published in a timely manner.
To that we say: “Well, sorry about that, but these predictions are still valid, even though it’s February!” Here is another in a list of articles eWEEK is now publishing about predictions for 2020 (and sometimes beyond!).
Dongyan Wang, VP of AI Transformation at Landing AI:AI adoption in non-consumer internet industries is still in an early stage, and many projects are stuck in pilot purgatory because of challenges ranging from a lack of data to knowing how to manage complex machine learning workflows. In 2020, we will see the emergence of end-to-end, verticalized AI platforms that will enable customers to move out of pilot purgatory and take their AI projects all the way to the finish line.

#responsive_asset {
border-bottom: none !important;
border:#2b9fea solid 3px;

Bruce Milne, Pivot 3:IT will begin recognizing the power of video-based computing: As video analytics improve in 2020, endless opportunities will emerge for IT. Video-based data currently makes up 60% of data collected, and in previous years has largely been considered by enterprises to be a liability and an expense. In 2020, we’ll begin to see a shift, with enterprises turning to video to optimize their product offerings or support strategic initiatives; for example, a city might secure its transit systems using video, but incorporate analytics to mine the data for insights like capacity needs.
Further reading

Matt Kunkel, CEO of LogicGate:RPA beats out AI in risk and compliance: There are a number of flavors of AI and RPA (robotic process automation) and machine learning. With regard to risk and compliance, the one that will continue to make inroads in 2020 is RPA. The reason: When needing to analyze large volumes of data for Fortune 500 companies, the volume of data just isn’t there to make the predictions of AI relevant. RPA works so well because many risk and compliance functions follow a formal process and there’s a much clearer path to automate those steps as companies put more and more data through the specific process. Then the question becomes how to optimize and iterate on that system. Other areas ripe for RPA application include third-party risk, IT, policy and procedure, and internal audits.
David Jones, VP of Marketing at AODocs:AI is not a silver bullet to solving content management: We tend to think of AI as a silver bullet, something that will resolve all business problems with just one algorithm implementation. However, this is a fallacy. We need to move away from the idea that one monstrosity of an AI algorithm will get this done, and instead shift to the idea of many AI bots working together to optimize previously stored data. In 2020, AI will be deployed against legacy databases to identify what data was stored in the first place, delete what is no longer needed and assign enriched metadata for better, more refined search and streamlined record-keeping. One behemoth won’t do this, but rather, a set of interconnected algorithms.
Cheryl Wiebe, Practice Lead, Industrial Intelligence Consulting at Teradata:—What the world is calling AI today will split into several areas in 2020, which someone in marketing will inevitably create pithier names for. These include RPA; automated feature engineering and selection; perception AI, which is the automation and refinement of physical perception; and resource allocation AI, the marriage of optimization technologies to sense and respond to demands in real-time.
—AI will begin to improve the process of data management itself. For example, for system resource allocation, for automated feature engineering, for operational metadata collection, and for better knowledge management (such as tagging).
Jeff Catlin, Lexalytics CEO: NLP and text analytics will become a bigger part of RPA solutions: Both Forrester and Gartner report that many RPA vendors are lagging behind in supporting trending text analytics use cases, lack capabilities with “unstructured document use cases” involving PDFs and have trouble fitting text analytics/NLP components into their larger environment. As companies automate larger and larger processes, NLP vendors offering viable solutions that meet RPA requirements—such as on-premises/hybrid cloud options, easy-to-integrate APIs, customizability and quick ROI—will rush in to fill the void.
Chad Meley, VP of Marketing at Teradata:—After a few successful AI pilots over the last couple years, enterprises will put a renewed focus on enterprise data management and integration, serving as a foundation to scale up to hundreds and thousands of narrowly defined AI use cases. Every sort of machine intelligence that surrounds us today is narrow AI. In 2020, a successful enterprise AI initiative will spawn hundreds if not thousands of use cases, each supported by a narrowly defined algorithm.
—There will be immense interest and adoption of “no-code analytics.” We’ve seen a steady democratization of advanced analytics by automating away certain laborious aspects such as feature engineering and model selection. But advanced analytics becomes truly pervasive when machine learning and other advanced procedural analytics becomes something that requires absolutely no coding or SQL skills. No-code analytics becomes embedded in workflows or invoked through simple drop-down menus. They won’t make coding obsolete in the analytics world, but will increase the number of use cases benefitting from analytics in large enterprises by a factor of 100.
Jeff Catlin, Lexalytics CEO: The biggest research advancements in AI will be theoretical: For the last five years applications of AI have run far ahead of our understanding of how this all works. With some big practical advances in the latter half of 2019, I predict we’re due for fewer world-beating algorithmic inventions and more progress on the theoretical side explaining why any of this works. The field has been moving quickly, so by the end of 2020 the balance will be shifting again with the theoretical work paving the way toward a new generation of algorithms.
Jeff Catlin, Lexalytics CEO: Less magic and more solutions: AI will have a good year and will solidify its position as the defining technology of the next decade. Providers seem to have wised up and are no longer pushing the Magical AI angle, and are instead pushing the correct message that AI can aid humans, making them faster and better at their jobs.
Muddu Sudhakar, CEO of will disrupt traditional IT/Cloud/DevOps: The core of DevOps is about improving agility and flexibility; AIOps offers the ability to help by automating key steps from development to production, predicting results in production, and automating response to changes in the production environment. Even as microservices, hybrid cloud, edge computing and IoT raise app complexity and increase the volume of logs to search for root causes, AIOps streamlines aggregating of data from multiple systems, while DevOps adds efficiency by integrating previously siloed systems. AIOps, like DevOps, spurs culture change because it requires looking system-wide rather than narrowing on specific technologies or infrastructure layers. It also requires a comfort level with high rates of automation.
Jeff Catlin, Lexalytics CEO: Self-driving … still a long way away: While AI in corporate settings will do really well, there will be a few spectacular failures of AI, most notably in the area of self-driving cars. Tesla’s new Smart Summon is quite impressive, but still has a way to go. Its widespread use by the Tesla community will result in lots of videos of slow-speed accidents where they run into cars, light poles and even people.