Survey finds majority of companies with AI projects reporting positive results. IT leaders see potential for security analytics and predictive intelligence to improve their delivery of tech services.
Artificial intelligence, seen as the cure-all for a plethora of enterprise shortfalls, from chatbots to better understanding customers to automating the flow of supply chains. However, it is delivering the most impressive results to information technology departments themselves, enhancing the performance of systems and making help desks more helpful. At the same time, there’s a recognition that AI efforts — and involvement — need to expand beyond the walls of IT across all parts of the enterprise.
This is one of the takeaways of a recent survey of 154 IT and business professionals at companies with at least one AI-related project in general production, conducted and published by ITPro Today, InformationWeek and Interop. Among those survey respondents with at least one AI application in general production, those with “excellent” and “very good” results comprise 64% of the group — excellent results account for 23% of respondents and 41% report very good results.
Looking at the characteristics of the successful AI leaders, top use operational cases include predictive maintenance (54%), Inventory and supply chain optimization (50%) and manufacturing analytics (50%). At the same time, many respondents see the greatest benefits going right to the IT organization itself — 63% say they hope to achieve greater efficiencies within IT operations. Another 45% aim for improved product support and customer experience. Another 29% seek improved cybersecurity systems.
The top IT use case is security analytics and predictive intelligence, cited by 71% of AI leaders. Another 56% say AI is helping with the help desk, while 54% have seen a positive impact on the productivity of their departments. “While critics say that the hype around AI-driven cybersecurity is overblown, clearly, IT departments are desperate to solve their cybersecurity problems, and, judging by this question in our survey, many of them are hoping AI will fill that need,” relates Sue Troy, author of the survey report. “On the help desk, meanwhile, AI tools are using predictive analytics to improve decision-making around incident management and demand planning. And AI is being used for help desk chatbots and intelligent search recommendations.”
There is a significant need for AI expertise and skills. More than two in three successful AI implementers, 67%, say they are seeing shortages of machine learning and data modeling skills, while 51% seek greater data engineering expertise. Another 42% say compute infrastructure skills are in short supply.
Security ranks as the top concern among successful AI implementers, with 44% citing this as their leading issue. Model transparency – or the degree to which the inner workings of AI algorithms are visible to users of the technology — was the second-leading concern, as cited by 36%, “Model transparency is an especially thorny issue,” Troy relates. “A high level of transparency can help mitigate bias and promote trust of the system, but it carries concerns that model explanations can be hacked, making the tech more vulnerable to attack.” Built-in bias follows among 33%, as well as concerns about unexpected or unusable outcomes with 33%.
When asked about specific AI technologies they expected to incorporate into their workplaces in the next six to 24 months, machine learning tops the list among successful AI sites, cited by 55%. Deep learning follows at 53%, and intelligent robotic process automation (RPA) rounds out the top three at 52%.
Successful AI projects take time to roll out. The typical AI project took six months to a year to complete, close to half of successful AI implementers (47%) indicate. Close to one-third, 32%, report taking more than year. Only 21% were able to wrap up AI initiatives in less than six months. The costs of these projects were kept in line — 45% said the project cost about as much as planned, while 25% said the costs ran over budget. By contrast, 40% of those with less-successful AI initiatives report cost overruns. “The more experienced IT practitioners are with AI, the better able they are to project costs and avoid going over budget,” Troy says.
Disclosure: Smita Nair Jain has nothing to disclose. She doesn’t own stock in any publicly traded companies and does not hold investments in the technology companies. She has equivalent of the American 401(k) plan in India that is automatically managed. (Updated: July 06, 2020)
Disclaimer: The opinions, beliefs and viewpoints expressed by the various author(s), publisher(s) and forum participant(s) on this web site do not necessarily reflect the opinions, beliefs and viewpoints of the @SmitaNairJain or official policies of the #SmitaNairJain
In This Article (Categories, Tags and Hashtags): #ArtificialIntelligence, #BeTheChange, #Data, #DataScience, #Diversity, #Economists, #Equity, #GenderEquality, #Inclusion, #LeanIn, #LeanInTogether, #MachineLearning, #MindfulLeadership, #Mindfulness, #OptionB, #Programming, #Visualization, #WhoRunTheWorld, #WomenEmpowerment, #WomenEntrepreneurs, #WomenHelpingWomen, #WomenInAG, #WomenInAI, #WomenInArtificialIntelligence, #WomenInBusiness, #WomenInConstruction, #WomenInDigital, #WomenInEconomics, #WomenInEngineering, #WomenInLeadership, #WomenInMath, #WomenInMedicine, #WomenInPower, #WomenInResearch, #WomenInSales, #WomenInScience, #WomenInSTEM, #WomenInTech, #WomenInTechnology, #WomenLeadTheWay, #WomenPower, #WomensAgenda, #WomenSupportingWomen, #WomenWhoCode, #WomenWhoLead