“Companies will adopt AI — not just because they can, but because they must,” IDC’s AI program vice president Ritu Jyoti noted. "AI is the technology that will help businesses to be agile, innovate, and scale.”
The arrival of AI capabilities in the enterprise is no longer theoretical. “The last year has demonstrated a rapid acceleration that has changed the question from ‘Where do artificial intelligence technologies fit within our organization?’ to ‘What areas are not yet right for AI?’” says Josh Perkins, field CTO at digital business consultancy AHEAD. “General applications across the enterprise continue to become increasingly evident by the day.”
[ Check out our primer on 10 key artificial intelligence terms for IT and business leaders: Cheat sheet: AI glossary. ]
AI will deliver massive value when intelligent tools and capabilities are used to solve industry-specific problems, says Perkins, noting that creative applications are emerging across enterprise organizations in healthcare, banking, insurance, retail, and manufacturing. “This is driven in no small part by a desire to better monetize data assets and utilize new data streams to unlock insights,” Perkins says.
5 areas where AI tools help with digital transformation challenges
As technology leaders plan their real-world AI-enabled digital initiatives, it’s helpful to understand where the most value exists. Certain themes are emerging consistently across sectors. Let’s examine some of the most functional use cases in production now across the spectrum of AI: from machine learning (ML) and natural language processing (NLP) to edge AI and AIOps.
1. Conversational AI: Revamping the customer service experience
What do you get when you combine rich customer behavior data, NLP, and chatbots? The potential to transform customer contact and support, often without human involvement.
“Dramatic improvements to NLP are making customer experiences richer and more dynamic each day,” Perkins says. “This technology is advancing the depth and natural flow of conversations between bots and the customer.”
This approach enhances customer self-service when it enables expedited access to backend systems - hopefully solving the customer issue faster. Within the next few years, Perkins predicts, it will be harder for customers to discern whether they are conversing with a bot or a human customer service agent.
Indeed, IDC determined that the deployment of automated customer service agents was the top AI use case based on worldwide spending in 2020. “There are many use cases currently being applied in the retail and e-commerce verticals, largely focused on customer service,” Perkins says. “Within healthcare, conversational AI is being used to assist with patient support and appointment scheduling.”
2. Edge AI: A cure for bandwidth, latency, and privacy problems
AI once lived solely in the domain of the world’s largest data centers. But as AI has moved to the outer edges of networks, it’s beginning to solve a host of distributed data and analytics problems for the enterprise. Edge AI is the embedding of intelligent capabilities at the point of data origin, whether that’s an IoT endpoint, a smartphone, or a connected car. “Put another way,” Red Hat chief technology strategist E.G. Nadhan explains, “edge computing brings the data and the compute closest to the point of interaction.”
[ Get a shareable primer: How to explain edge computing in plain English.]
Edge AI has been rapidly expanding, powering everything from smart speakers to street-corner cameras. Now it’s becoming a boon for the enterprise.
“Until quite recently, AI at the edge remained largely theoretical, but in 2021 we are likely to see a rise in AI processing in a wide span of edge products due to advancing technologies which are both more accessible and affordable,” says Orr Danon, CEO of Hailo, a maker of edge processors. “AI at the edge will be essential to managing burgeoning amounts of data and alleviating the growing strain on business networks. Processing data right at the edge without the need to transfer it to the cloud makes devices more powerful, versatile, responsive, and secure – and aids in regulatory compliance.”
Retailers can deploy AI at the edge to process in-store video locally, quickly, and with minimal latency, laying the groundwork for touchless, cashier-less shopping in some cases. Stores can use cameras and edge AI to detect objects from a distance and process the related information swiftly. That data can help with customer wait times, stocked shelves, and the overall in-store experience.
Likewise, manufacturers can apply edge AI to process production line data. This could help to complete quality inspections or to enforce social distancing or other employee safety measures.
[ Want to learn more about implementing edge computing? Read the blog: How to implement edge infrastructure in a maintainable and scalable way. ]
3. Machine and deep learning: Fighting fire with fire in cybersecurity
Bad actors are already harnessing AI to launch sophisticated phishing attacks and other insidious cyber assaults, using intelligent automation to increase the velocity, volume, and variety of their attacks. Deep fake scams, for example — which Forrester predicts will cost organizations more than $250 million this year — employ AI to create convincing audio and video to trick users in their email compromise attacks.
Traditional cyber mitigation techniques are no match for such sophisticated approaches. Thus the use of AI in cybersecurity and attacks was one of Gartner’s top nine security trends for 2020, noting the necessary AI enhancement of cybersecurity defenses.
There is a multitude of AI cybersecurity applications in cybersecurity and threat intelligence.
The most common use cases involve facial and speech recognition, spam or phishing identification, and malware detection. Machine learning methods can be applied to detect anomalies in email, pattern recognition technology can identify regulated personal data in need of protection, unsupervised machine learning can categorize websites and identify high-risk sites, and unsupervised machine learning can suss out near-duplicates in phishing and spam attempts. A recent TrendMicro article points to the use of end-to-end deep learning as a solution for detecting malware.
4. AIOps: Relief for IT alert fatigue and more
Do IT organizations need yet another -OPs acronym to think about? In a word: yes. IDC pointed to IT automation as one of the fastest-growing use cases for AI in 2020 (along with pharmaceutical R&D and HR automation). As Eveline Oehrlich, chief research director at DevOps Institute, pointed out in a recent article, AIOps can prove transformative for IT organizations in which the environment generates so much data that decision making has suffered. This is an ever-growing cohort of IT functions in the era of hybrid cloud. ML can save the day for IT teams by addressing large volumes of often redundant alerts, assisting in managing systems performance in a more real-time or proactive way, and delivering greater end-to-end visibility.
We included AIOps on our list of 10 top AI trends in 2021 for good reason. Siloed monitoring systems can’t keep up with today’s variegated environments. Gartner sees five primary use cases for AIOps: performance analysis, anomaly detection, event correlation and analysis, and IT service management.
“Together, these build a comprehensive layer of production and operational insights analysis that can run on big data and against advanced modern software architecture,” Eran Kinsbruner, chief evangelist and product manager at Perfecto by Perforce, wrote in this recent article. “With the power of AI-based operations, teams can focus on determining the service health of their applications and gain control and visibility over their production data.”
As vendors begin to offer AIOps platform solutions, Forrester advises that IT leaders look for those that can deliver cross-team collaboration capabilities, end-to-end digital experiences, and seamless integration into the whole IT operations management toolchain.
5. Machine learning: Predictive resource optimization
The value of being able to predict sudden shifts – in supply or demand, in healthcare outcomes, in sales or customer behavior – was never clearer than in 2020.
Explained at a basic level, supervised machine learning – in particular, regression – enables organizations to build mathematical models to predict future outcomes based on a series of predictor variables or inputs. “The business applications of this approach are abundant across industries,” says Perkins. The common denominator is the ability to do more with less.
MORE ON ARTIFICIAL INTELLIGENCE
“Whether the resources are human, inventories, or discreet processes, machine learning lets us observe and define patterns to gain previously unreachable insight,” Perkins says. Use cases of this technique include inventory optimization and reorder points, proper scheduling of employees to work during a specific shift or periods of demand, and even increasing the accuracy of sales forecasting.
[ Get the eBook: Top considerations for building a production-ready AI/ML environment. ]