If you are currently in the market for almost any kind of enterprise software, you will almost certainly run across at least one vendor claiming that its product includes artificial intelligence (AI) capabilities. Of course, some of these claims are no more than marketing hyperbole, or "AI washing." However, in many cases, software makers truly are integrating new capabilities related to analytics, vision, natural language, or other areas that deserve the AI label.
The market researchers at IDC have gone so far as to call AI "inescapable." According to the firm, "By 2025, at least 90% of new enterprise apps will embed artificial intelligence."
Similarly, Omdia | Tractica predicted that worldwide revenue from AI software will climb from $10.1 billion in 2018 to $126.0 billion in 2025, led in large part by advancements in deep learning technology. It added, "Tractica believes the global market has entered a new era where AI is viewed as an essential technology to driving improvements in quality, efficiency, and speed."
IDC's Frank Gens stated the same idea even more strongly in the IDC report. "It's hard to overstate the importance and the impact that artificial intelligence will have on enterprises' ability to create new products and services, new customer experiences, and new ways of operating in the coming decades," he said. "By 2025, we expect to see enterprises using AI-enabled and AI-led apps to gain competitive advantage from shorter reaction times, greater success with product innovation, and improved customer satisfaction."
That perception of the importance of AI drives many enterprise software buyers to look for applications that incorporate AI capabilities. And of course, vendors have quickly obliged.
In some cases, vendors have incorporated machine learning-based capabilities into existing software. In other cases, startups are taking an "AI-first" approach, developing entirely new categories of enterprise software designed from the very beginning on AI technology. In both cases, the AI features are enabling enterprises to do new things in new ways.
The following slides highlight 10 ways that artificial intelligence is currently transforming enterprise software and gives hints about where enterprise use of AI might be headed next.
Streamlining data preparation
AI and analytics projects all have one thing in common: voracious appetites for data. Processing and cleaning that data so that it can feed those analytics and AI systems is a huge job — a job that is now being made simpler by AI. Vendors like Unifi Software (now owned by Boomi), Paxata (now owned by DataRobot), Alation, Trifacta, Io-Tahoe, and others offer data preparation and/or data cataloging tools that both rely on and power machine learning. Tools like these often provide the first step toward implementing AI in other areas of the enterprise.
Securing enterprise networks
Cyberattackers and enterprise security teams are constantly looking for new ways to outsmart one another. Some of the newer tools in the enterprise arsenal are applications that integrate machine learning to detect brand new types of threats in real time. Vendors include companies like Darktrace, BlackBerry Cylance, Tessian, and others. Many of these tools use machine learning techniques to spot anomalies in network traffic, email, or user activities. As a result, they are able to identify when an attack is taking place — and potentially halt or mitigate the attack — even if the attack is unlike any the organization has seen before.
Transforming IT operations through AIOps
You've undoubtedly heard of DevOps. You may also have heard of some of its variants, such as DevSecOps, NoOps, NetOps, or DataOps. AIOps is one of the latest variations on the DevOps themes, and it has the potential to be one of the most transformative.
While DevOps embraces automation to help simplify IT operations, AIOps goes a step farther. It runs machine learning algorithms on IT operations data in order to surface insights that optimize and enhance those operations. In some cases, they can also automatically take action based on those insights, allowing IT personnel to oversee much larger IT environments than would otherwise be possible. Companies that sell AIOps solutions include BMC, Splunk, IBM, Broadcom.
Processing extremely complex documents
Enterprises have been using optical character recognition (OCR) systems to extract data from templated documents for decades. In the past, these symptoms only worked with documents that followed a rigid format. Today's more advanced AI-systems can look at any document and figure out where the important information is on the page. For example, you could feed these systems invoices from a wide variety of different vendors, all with different layouts, and the AI could figure out who the invoice was from, the amount to be paid, the due date, etc. Companies with products in this space include Element AI, Google Cloud, UiPath, and others.
Understanding natural language in conversation
You've heard about (and probably interacted with) chatbots before, so natural language processing (NLP) capabilities may seem like nothing new. However, vendors are continuing to improve their NLP capabilities and apply them to new use cases that allow enterprises to improve their customer service. For example, communications software vendor Twilio has an application called Media Streams that uses AI to analyze customer service calls in real time. It not only transcribes the call, it uses NLP to understand the content of the call and bring up related knowledgebase articles that might help agents solve the problem. It can even authenticate customers based on their voice and analyze their emotional state.
Analyzing the business using pre-trained models
Critics of AI point out (rightly) that AI projects are incredibly complex and frequently don't deliver the immediate results that business leaders expect. The nature of machine learning requires building a model and then training that model on large sets of data. And if you want to be successful with this process, you'll need clearly defined goals, a highly skilled data scientist, lots and lots of data, and plenty of time and patience.
One way to speed up this process and simplify AI-based analytics is by using pre-trained models and AI accelerators. A growing number vendors, including Outlier, have begun offering specialized products for specific industries or use cases.
Some observers believe the area of AI-based testing is poised to take off in the near future. One such tool that is already on the market is Applitools. It allows organizations to dramatically streamline their testing of user interface components without having to write scripts, in some cases reducing testing time by 75%. And its computer vision tools can easily spot differences between versions of an interface. Other vendors that offer testing tools that leverage machine learning and other AI capabilities include Tricentis, Cigniti, Sealights, Test.ai, and more. Many are cloud-based and integrate into existing DevOps or Agile workflows.
Optimizing supply chain
The shortages associated with the 2020 pandemic left everyone more aware than ever of the critical importance of supply chain management. A small but growing number of enterprises are now investing in supply chain management applications with AI features that can help them improve just-in-time deliveries, reduce costs, anticipate potential problems, and recover from disruptions. Those investments appear to be paying off as a study by McKinsey found that among organizations that were using AI for supply chain management, 61% experienced cost decreases and 63% saw revenue increases. Well-established technology giants like Oracle and SAP are beginning to integrate AI and machine learning into their applications, and some enterprises are also using general-purpose business intelligence or analytics platforms with AI capabilities for similar purposes.
Enabling rapid innovation
In addition to disrupting supply chains, the coronavirus pandemic has also forced enterprises to innovate quickly in order to accommodate new ways of working, adapt to new regulations, and take advantage of new opportunities. For example, FinTech vendor Kabbage, which offers cash-flow solutions for businesses, used AI to capitalize on the opportunity offered by the Paycheck Protection Program (PPP). During the three months when the PPP was active, Kabbage's automated systems allowed the company to process and approve 209,000 loans, making it the third-largest PPP lender. That feat was only possible for such a small company because its AI systems allowed it to process 75% of loan applications without human intervention.
Applying computer vision to new problems
Many news stories have covered computer vision research into areas like medical imaging analysis and self-driving cars, but startups are also applying computer vision to some tasks you might not have considered. For example, vendor Tractable makes AI-based tools that can help insurance agents assess damage to vehicles, homes, and businesses. It can provide a check on the claims process, flagging estimates that seem out of line with best practices, or even automate the entire damage estimation process to provide faster customer service.
As AI continues to evolve and improve, look for even more diverse types of enterprise software to integrate artificial intelligence capabilities like these.
Cynthia Harvey is a freelance writer and editor based in the Detroit area. She has been covering the technology industry for more than fifteen years. View Full Bio