As artificial intelligence (AI) continues to evolve, its influence on decision-making processes in organizations grows exponentially. To better understand how AI is impacting organizational abilities to improve business outcomes, Teradata sponsored an IDC survey of 900 organizations across the globe.
We discussed decision-making and AI with Dan Vesset, group vice president at IDC, co-author of the IDC InfoBrief “Leadership and Decision Making in the Era of Generative AI,” published November 2023 and sponsored by Teradata, that gathers the survey findings; and guest speaker at the corresponding webinar.
TERADATA: What was the purpose of the AI-driven decision-making study?
DAN VESSET: Commercial availability of generative AI awakened the imaginations of millions across companies, government agencies, and nonprofit institutions. At the end of 2023, it’s safe to say we witnessed a tipping point in the adoption of all types of AI. As excitement about AI turned to a more pragmatic focus on how to deploy and scale AI to augment and automate decision-making processes, we started noticing an increased focus on investments in data to ensure AI models have access to large volumes of trusted data. The goal of our study of 900 organizations across the globe—focusing on their chief data officers (CDOs), chief information officers (CIOs), and chief financial officers (CFOs)—was to better understand how AI is impacting organizational abilities to improve business outcomes and how organizations at different levels of maturity are investing and benefiting from data and AI capabilities, including how different personas collaborate to achieve desired goals.
TERADATA: How do you define AI-driven decision-making?
VESSET: AI-driven decision-making is the capability to turn experience-based or “gut feeling” decision-making processes into those that leverage large language models (LLMs) and AI trained on firm-specific data. Decision-making in this case is not simply the ability to organize and store large amounts of data, develop dashboards, or even train new AI models. It’s leveraging these and other related capabilities to treat decision-making as a process consisting of multiple tasks. Each task needs to be evaluated for the opportunity to automate it with AI and the role of the human subject matter expert in governing that AI-driven automation.
TERADATA: What impact do organizations with high levels of enterprise intelligence see?
VESSET: If I had to choose one factor: scalability. But there’s a lot packed into that word. It’s the ability to:
- Start small with a few decision-making processes and scale them across business functions
- Identify similar patterns of AI and data processing and leverage technical and AI expertise across use cases
- Extract value by freeing up employees to focus on other activities, such as those that require close interpersonal relationships or those that can’t be automated
- Use AI and data to standardize tactical or operational decisions
- Standardize governance practices or guardrails for AI to ensure the highest possible acceptance levels of AI-generated recommendations
TERADATA: How prepared are leaders to use generative AI?
VESSET: It will be important to monitor organizational preparedness to use generative AI over the short term. The vast opportunities will focus leadership to take a detailed look at current technology, data, and skills availability and assess the need for additional investment in these resources. Some organizations are establishing partnerships with preferred technology providers that share their vision for AI. Others are ensuring availability of data to train AI models. We’re also seeing leading organizations creating collaborative teams with representation from AI, data, application development, and business to ensure resources are allocated to use cases with the greatest opportunities to generate a return in the short term.
TERADATA: What was the most surprising finding from the survey?
VESSET: Leading organizations were more likely than other organizations to have AI staff distributed across business functions rather than in a central service or center of excellence (COE). I think it points to the need to bring AI expertise close to business expertise to enable rapid experimentation, iteration, and scaling of AI-powered decision-making.
TERADATA: How can organizations improve data literacy for generative AI?
VESSET: Organizations should put training in place to explain to all employees the potential risks with currently available generative AI technology. Training should be continuously updated as generative AI solutions evolve. Besides training, improve trust in the underlying data that trains generative AI with initiatives for data governance, data lineage, data cataloging, data quality, and data observability.
Discover more insights from Vesset in the IDC InfoBrief and on-demand webinar.