AI Is The Uncertainty Cure Enterprises Want In 2020

• 85% of enterprises are evaluating or using artificial intelligence in production today.
• 55% of all enterprises adopting AI today are using TensorFlow as their primary development tool.
• 73% of enterprises with the most advanced AI adoption levels say supervised learning is the most popular machine learning technique (73%).
• Human-in-the-loop AI models are considerably more popular among enterprises with advanced AI expertise compared to their peers.
• Enterprise's enthusiasm for AI is growing, with 62% increasing their spending last year, according to a recent MIT Sloan Management Review study.

These and many other insights are from O'Reilly's recently published research, AI Adoption in the Enterprise 2020 available for download here (20 pp., PDF, free, opt-in). The survey is based on interviews with 1,388 respondents from 25 industries, with 17% of total respondents from the software industry. 30% of respondents are data scientists, data engineers, AIOps engineers, or their managers. 70% of all respondents are in technology roles. For additional information on the methodology, please see pages 2 and 3 of the study, downloadable here.

Additional insights from the study showing enterprises' growing adoption of AI include the following:

• In 2020, AI is being adopted evenly across enterprises, with R&D leading all departments by a wide margin. O'Reilly's survey finds that enterprises are stabilizing their adoption patterns for AI across a wide variety of functional areas. Nine to twelve functional areas included in the survey have over 10% adoption. It's fascinating to watch how IT is adopting AI to improve ITSM performance for example. AI has the potential for redefining enterprises, making them more customer-driven, adaptive, and capable of generating and sharing intelligence faster than ever before. Guiding the transformation of an enterprise takes a framework that both can create knowledge while staying focused on the customer and provides an understanding of the entire customer journey. BMC's Autonomous Digital Enterprise (ADE) shows potential in this area, as its structure enables all departments of an enterprise to contribute and share AI-driven insights about customers and provide a transcendent customer experience. The following is a ranking of the functional areas of enterprises where AI is used today:

• Supervised machine learning algorithms is the most popular machine learning technique in enterprises today. 73% of enterprises with advanced expertise in AI are making extensive use of supervised machine learning techniques to interpret, classify, and analyze the large data sets they've accumulated over years of operations. Enterprises who are evaluating AI are using deep learning techniques in their pilots, making this area of machine learning most popular with enterprises running AI pilots today.

• TensorFlow continues to be the most popular development tool across all enterprises evaluating and using AI in production today. The O'Reilly research team found that 55% of all enterprises in 2019 and 2020 place a high priority on TensorFlow expertise, making it the most popular tool two years in a row. TensorFlow is integral to completing deep learning and neural network projects, further evidence of how enterprises are adopting AI to solve increasingly complex problems. The following is an analysis of the AI tools enterprises are using today: 

• 53% of advanced enterprises using AI today say the greatest risk when building and deploying Machine Learning models is unexpected outcomes and predictions. The more experienced an enterprise is using AI, the more they're likely to anticipate unexpected outcomes and work on making models more transparent. The most advanced enterprises adopting AI today are also far more likely to include steps during model building to improve fairness, ethics, and limit or control biases. 

• Of the 15% of enterprises who are considering AI, one in five or 22% says that a lack of institutional support is slowing down adoption efforts. The most significant barrier to overcome is changing a company culture that doesn't recognize the value of AI. Conversely, the most successful AI implementations are known for the strong support for senior management they receive. Closing the skills gap is the third greatest impediment to making progress with AI. Comparing bottlenecks holding back AI adoption across the entire survey sample versus enterprises who have reached a level of AI maturity shows how significant the skills gap continues to be. The O'Reilly research team found that selecting the right machine learning technique for the job has more than three-quarters (78%) of respondents selecting at least two of ML techniques, 59%, using at least three, and 39% choosing at least four.

Sources:

10 Ways AI Is Improving Manufacturing In 2020, Forbes, May 18, 2020

AI Is Transforming The Enterprise, KMPG, 2020

Autonomous Digital Enterprise Executive Brief, BMC, 2020

Capitalizing on the promise of artificial intelligence, Deloitte Insights, 2020

Enterprises Increased AI Spending By 62% Last Year, Forbes, March 15, 2020

How COVID-19 Is Changing Analytics Spending, Forbes, May 10, 2020

Industry's fast-mover advantage: Enterprise value from digital factories, McKinsey & Company, January 10, 2020

Roundup Of Machine Learning Forecasts And Market Estimates, 2020, Forbes, January 19, 2020

State of AI in the Enterprise, Deloitte Insights, 2020

The Rise of the AI-Powered Company in the Postcrisis World, Boston Consulting Group, April 2, 2020

Top 8 Data Science Use Cases in Manufacturing, ActiveWizards: A Machine Learning Company Igor Bobriakov, March 12, 2019