According to respondents in a recent global survey, data for machine learning must be AI-ready, and talent shortages are an impediment
In a May–June 2022 global survey on bridging the gap between Business Intelligence and AI involving 600 CIOs across 18 countries 14 industries comprising organizations earning US$500m or more in annual revenue, AI and data management were cited as essential pillars to enterprise success, but data mismanagement could jeopardize respondents’ future AI success.
Besides CIOs, the respondents comprised C-level data/analytics officers and other senior technology executives (evenly distributed among North America, Europe, and the Asia-Pacific pacific region).
For the large firms in the Asia Pacific region, findings include:
- 41% of respondents indicated expectations that AI will play a critical part in the finance function by 2025, with 30% citing better security and risk management as the biggest benefit of AI/ML adoption.
- 39% cited expectations for the period up to 2025 that investments in talent and skills development were a top priority, followed by 29% citing developing ML infrastructure as a top priority, in scaling benefits from the use of AI and ML.
- 74% indicated that scaling AI to create business value was their top priority for enterprise data strategy by 2025; 78% indicated that problems with data could be the biggest factor that affects AI and ML goals between now and 2025.
- 32% of respondents in the region cited rigidity of organizational structures; constraints on budgets for new tech as top impediments to AI/ML adoption, while 31% cited shortage of AL and ML qualified talent.
- 72% of respondents in the region indicated their belief that multi-cloud ensured the most flexible foundation for AI development.
According to Chris D’Agostino, Global Field CTO, Databricks, which commissioned the survey and report: “These insights are consistent with what we hear in the field. AI-ready data is no longer a nice-to-have— it is critical to solve real-world problems and drive business outcomes,” alluding to the use of an open and unified data platform to let organizations put their data into action in mission-critical AI projects.
Besides data issues, organizations need to improve processing speeds, governance, quality of data and “sufficiency for models”, to achieve and scale their AI goals, according to the report.