More than half of enterprises are in the “mature” phase of AI adoption – defined by those currently using AI for analysis or in production – while about one third are evaluating AI, and 15% report not doing anything with AI, an O’Reilly survey reveals.
These numbers demonstrate growth when compared with O’Reilly’s 2019 report, which found just 27% of organizations in the “mature” adoption phase and 54% in the evaluation phase.
Organizations to institute formal data governance
When it comes to data governance, more than 26% of respondents say their organizations plan to institute formal data governance processes and/or tools by 2021 and nearly 35% expect this to happen in the next three years.
Currently, just one-fifth of respondent organizations report having formal data governance processes and/or tools to support and complement their AI projects, similar to findings uncovered in the survey.
Orgs struggling to expand and scale AI practices
Difficulties in hiring and retaining people with AI skills was once again noted as a top barrier to AI adoption in the enterprise, down slightly from 18% in 2019. As in 2019, the biggest bottleneck to AI adoption was reported to be a lack of institutional support (22%), followed by “Difficulties in identifying appropriate business use cases” at 20%.
“AI practices are maturing, and adopters are experimenting with sophisticated AI techniques and tools, which bodes well for the future advancement of AI in the enterprise,” said Rachel Roumeliotis, O’Reilly Strata Data & AI conference co-chair and strategic content director at O’Reilly.
“However, organizations will continue to struggle to expand and scale their AI practices if they don’t address the importance of data governance and data conditioning in ML and AI development.”
By a 2:1 margin, respondents in companies that are evaluating AI cited an unsupportive culture as the primary bulwark to AI adoption, suggesting increased resistance for organizations who have yet to put AI into production. By contrast, AI adopters are about one-third more likely to cite problems with missing or inconsistent data as the biggest bottleneck.
AI efforts: Other findings
- Among mature adopters, supervised learning was reported to be the most popular machine learning technique (73%), while deep learning (55%) is the most popular among organizations still in the evaluation stage of AI.
- The bulk of AI use is in research and development—cited by just under half of all respondents—followed by IT, which was cited by just over one-third. Another high-use functional area of AI is customer service, with just under 30% of share.
- Respondents identified the most critical ML- and AI-specific skills gaps in their organizations as the shortage of ML modelers and data scientists (58%), almost exactly on par with findings in 2019. This was followed by the challenge of understanding and maintaining a set of business use cases (49%) and data engineering (40%).
- Unexpected outcomes/predictions were the single most common risk factor when building and deploying ML models, cited by close to two-thirds of mature—and by about 53% of still-evaluating—AI practitioners.
- TensorFlow remains the most popular tool for use in AI-related work, as reported by roughly 55% of respondents in both 2019 and 2020. Additionally, four of the five most popular tools for AI-related work are either Python-based or dominated by Python tools, libraries, patterns, and projects.