10 Characteristics of an AI-Powered Enterprise
Artificial intelligence is seeping into enterprises in various ways. Smart enterprises integrate AI into their business strategy.
![AI AI](https://eu-images.contentstack.com/v3/assets/blt69509c9116440be8/bltd275a804ad078110/64bf3f2b13221c9e8ef082f1/00artificialintelligence-adrian_ilie825-adobe-cp.jpg?width=700&auto=webp&quality=80&disable=upscale)
adrian_ilie825 via Adobe Stock
Mainstream AI adoption is well under way. Some companies have an official AI strategy in place while others don't. Either way, everyday enterprise applications and tools are evolving to include AI. Whether enterprises realize it or not, they have an AI strategy by design or default. The former is a wiser course of action over the long term.
Some organizations have a separate AI initiative that's decoupled from the business strategy. When that happens, the outcomes may be fascinating lab experiments that aren't practical or scalable enough for real-world use.
Digital transformation makes the inclusion of AI as part of the business strategy even more important than it would be otherwise because digital organizations are software companies. Since commercial applications and tools are increasingly taking advantage of AI, the logical development by extension is AI embedded in enterprise-built applications. After all, businesses are moving more data and compute to the cloud and their new applications are being designed as cloud-first applications. Of course, AI and machine learning tooling is also available in the cloud, so developers have what they need to build “intelligent” applications.
AI and machine learning don't just work, however. They require testing and monitoring.
“Losing trust in AI-infused applications is a high risk for AI-based innovation,” said Diego Lo Giudice, VP and principal analyst at Forrester, in a blog post. “Forrester Analytics data shows that 73% of enterprises claim to be adopting AI for building new solutions in 2021, up from 68% in 2020, and testing those AI-infused applications becomes even more critical.”
Trust and safety are things that need to be proven through testing. As with software testing, some of the same questions arise that have to do with whether the testing is detailed enough, comprehensive enough, and executed by the right parties.
In short, enterprises should think strategically about their current and future requirements and ensure that those requirements and business goals are reflected in the company's implementation of AI.
Following are a few other traits of AI-driven enterprises.
Companies experienced with AI realize that narrow AI is the state of the art whereas less sophisticated organizations may think there's a general AI that simply needs to be pointed at data to produce a desired outcome. As vertical applications such as agriculture and financial services indicate, narrow applications of AI can be very effective.
“AI is not a magician with a powerful wand. It will not change our world overnight,” says Igor Ikonnikov, analyst and research & advisory director in the data & analytics practice at Info-Tech Research Group. “It's a highly automated and adaptable data processing unit.”
Another indicator of early maturity is the propensity to look at AI from a technical point of view. For example, if one inquires about a use case, the answer may come pack in the form of GPUs, neural networks, and Tensorflow.
“Identify where you can benefit from AI, where you have a lot of data to process, [and] where you have a lot of complexity. Always start by discussing the business scenario,” says Info-Tech's Ikonnikov.
In fact, Info-Tech says it does three things to help demystify AI for its clients, which are:
Explaining the different types of AI and their corresponding use cases so clients can consider AI without unnecessary agitation [about it being an existential threat] or unreasonable expectations about what it can do. InfoTech also helps identify the implementation opportunity that is most appropriate for the organization.
Cautioning clients about the common pitfalls associated with AI implementation so they can mitigate AI bias risks and maintain positive ROI from the implementation.
Suggesting an implementation roadmap with specific and manageable tasks so they can more easily get started with AI before bringing it in. Info-Tech also helps organizations improve their data management capabilities to prepare for a “data-insatiable” AI and to make the AI-empowered solution an integral part of the organization.
Multinational professional services network PwC uses AI throughout the company to reimagine and improve how the firm works and how it delivers work to clients. The firm created a platform called “Digital on Demand”, which enables its employees to practice and apply the AI knowledge they've gained as part of PwC's $3 billion upskilling program. Apparently, employees can develop specific tools with the technology to improve their everyday work. So far, more than 7 million hours of work have been automated.
“We’ve used AI to do some incredible work at PwC,” says Anand Rao, global artificial intelligence leader at PwC. “One of the reasons I believe it’s been so successful is our upskilling program, but also because we laid the groundwork for responsible AI before we started implementing AI more widely at PwC. Our responsible AI framework is incorporated everywhere to mitigate AI risks and biases. It's imperative that companies prioritize responsible use of AI to ensure fairness and detect bias.”
The “garbage in, garbage out” concept applies to AI. If the data isn't trustworthy and of high quality, the outputs are likely unreliable. According to a Capgemini report, “Improving data quality ranks as the number-one approach that AI-at-scale leaders use to get more benefits from their AI systems.”
In Capgemini's survey, which was the basis of the report, 94% of the AI-at-scale leaders are more likely to have achieved benefits that met or exceeded their expectations compared to 59% of struggling organizations.
Should AI be centralized or decentralized? How about a mix of both?
“The most successful digital enterprises have created Centers of Excellence where they are strategic about digital transformation that can be duplicated across the entire organization and not just digitizing processes for digitization’s sake [in] siloed departments,” says Anthony Macciola, chief innovation officer at digital intelligence company ABBYY. “It’s no longer the wild West where teams used AI for automation because they could. Now, they’ll use AI to identify the bottlenecks that are impeding the outcomes they desire and prioritize where and how to spend intelligent automation investments.”
Models tend to “drift” over time, becoming less accurate as new data is added or something else changes. Sometimes, the change is instant and dramatic. Other times, it may be more subtle.
Models need to be monitored so they can be tuned, retrained, or replaced as necessary. Expect to hear more in the coming months and years about the need to prioritize AI testing and monitoring.
2020 and 2021 have both proven that supply chains are not infallible, whether caused by unforeseeable spikes and dips in demand, ransomware attacks, erroneous assumptions, or the disruption of a major shipping route. Increasingly, enterprises in the supply chain and logistics industry are relying on AI and data insights to understand the vast number of external variables that can impact logistics success.
“Artificial intelligence allows third-party logistics providers to respond to shipping and delivery issues in real-time -- in some cases, having that response fully automated, [is] an ever-increasing priority for customers,” says Russ Felker, Chief Technology Officer at third-party logistics provider GlobalTranz Enterprises. “This approach creates more visibility for our customers and the end consumer and enables us to detect situations that require a customer to pivot based on variables like raw material shortages, increasing delivery times or condensed capacity within a certain market.”
AI, applied wisely, can help reduce operating costs. For example, GlobalTranz's technology is also used for cost predictions based on an analysis of data from multiple market sources, which is correlated with large internal datasets. In addition, GlobalTranz and the MIT Center for Transportation cooperatively developed a solution that describes how shipments can be optimally paired and consolidated to reduce total network transportation costs and efficiencies.
“Artificial intelligence will continue to serve as a useful tool for business intelligence and predicting industry trends, no matter the type of industry or nature of challenge,” says GlobeTranz's Felker. “The past couple of years have presented tremendous challenges across supply chains for all types of businesses. Enterprises that understand how to leverage AI will be a step ahead in tomorrow’s market.”
The pandemic has had a negative effect on mental health. Skylyte, which raises awareness of burnout, developed a proprietary model that examines hundreds of factors affecting personal resilience and burnout risk.
“Our machine learning models use a combination of self-reported and unobtrusive data to measure, understand and build team health and long-term productivity,” says SkyLyte CTO Alberto Escarlate. “After working with us, teams from various industries, including healthcare and tech, have seen a reduction of 47% in severe burnout cases, which is very significant because burned out employees are 260% more likely to churn.”
Global AI-powered identity verification service provider Shufti Pro uses AI to ensure rapid customer onboarding. Instead of requiring customers to spend an unnecessary amount of time, waiting in line to prove their identity, they can simply get their ID card verified through their webcam with a corroborating selfie, right in the comfort of their own home.
“By utilizing AI for customer onboarding, we have enabled banks and financial institutions to open accounts online and verify their customer’s identities remotely,” says Shahid Hanif, CTO at Shufti Pro. “[The] AI models can detect biometric fraud, identity theft and fraudulent documents in an instant, enabling businesses worldwide to minimize fraud attempts during the initial stage of onboarding -- before any long-term financial damage is done.”
Check out other InformationWeek slideshows.
Global AI-powered identity verification service provider Shufti Pro uses AI to ensure rapid customer onboarding. Instead of requiring customers to spend an unnecessary amount of time, waiting in line to prove their identity, they can simply get their ID card verified through their webcam with a corroborating selfie, right in the comfort of their own home.
“By utilizing AI for customer onboarding, we have enabled banks and financial institutions to open accounts online and verify their customer’s identities remotely,” says Shahid Hanif, CTO at Shufti Pro. “[The] AI models can detect biometric fraud, identity theft and fraudulent documents in an instant, enabling businesses worldwide to minimize fraud attempts during the initial stage of onboarding -- before any long-term financial damage is done.”
Check out other InformationWeek slideshows.
-
About the Author(s)
You May Also Like