Streamlining Processes Across Teams
Oracle has implemented ChatGPT Work and Codex across various teams, significantly reducing the time required for tasks that previously demanded specialist expertise. The company reports 130,000 active ChatGPT users and over 95,000 active Codex users, achieving a 98% decrease in the time needed for talent acquisition research.
Enhancing Recruitment Efficiency
The talent acquisition team developed a talent market intelligence tool utilizing ChatGPT Work. This tool converts job descriptions into a comprehensive set of market comparisons, compensation benchmarks, and assessments of the talent pool across relevant locations. Jan Ackerman, Senior Vice President and Global Head of Talent Acquisition at Oracle, stated, “We’ve gone from zero to a hundred,” noting that recruiters can now prepare for discussions in about 15 to 20 minutes, a process that previously took two to four days. The tool also standardizes the intake process, ensuring that every hiring manager receives consistent data and insights, regardless of which recruiter is involved.
Improving Business Workflows with Codex
The Oracle Applications Lab has created an ontology of internal objects, relationships, and rules, enabling Codex to translate plain-language inquiries into reliable SQL queries and manage calls to internal systems. One business user reported that a request that would have taken hours to fulfill now receives an almost immediate response from the new tool, with results matching those from the manual process. In site reliability engineering, Codex assists in gathering context about incidents and retrieving the appropriate playbook. Richard Lam, Group Vice President of the Oracle Applications Lab, mentioned that a typical incident resolution that used to take an hour can now be completed in minutes, although he emphasized the importance of human oversight in verifying systems and outputs.
Operational Insights and Future Directions
Oracle has identified several governance lessons from this transition: the necessity of establishing appropriate guardrails, favoring prototypes over formal specifications, and ensuring teams take ownership of the generated code. Lam cautioned about the importance of architecture, security, and maintainability, while Barry Shilmover, Vice President and Technical Advisor to the CIO, noted that teams are now able to prototype ideas directly with the tools. These changes are reshaping workflows across recruitment, analytics, and engineering, with widespread internal adoption expected to continue as employees leverage Codex and ChatGPT Work for emerging challenges.
Original source: OpenAI News