Oracle has spent billions helping other companies run AI, but its own companywide rollout came much later, and it’s now running into a familiar problem: code gets written faster while products don’t ship any sooner. At an internal town hall, co-CEO Clay Magouyrk told employees the company hadn’t found a way to make generative AI broadly useful to its own workforce a year ago, Business Insider reported.
“When I think about where we were a year ago,” Magouyrk said, “I don’t think we figured out how do we make AI really that useful for ourselves.” Oracle had made some progress using AI in customer support, but not yet across developers, finance and sales.
80% adoption in three months
That changed in April and May, when Oracle rolled out ChatGPT Enterprise and OpenAI’s Codex, according to Chief Information Officer Jae Evans. Rather than just making the tools available, the company set corporate standards, security controls and internal policies, and reached 80% adoption within three months.
Adoption brought costs. “We made it so easy to use and adopt that we might have gotten a little bit of sticker shock,” Evans said. Oracle now shows employees which models they’re using and what they cost. Evans noted that OpenAI’s GPT-6 Astra costs 2.5 times more than other models and pointed to lower-cost options for routine tasks.
The bottleneck moved downstream
“We see developers able to generate code using this tool in like a week’s time,” Evans said. “Normally, that would have taken a team of developers two to three quarters to go develop.” That speed hasn’t yet translated into getting products to customers faster.
“When you make the actual act of writing the code quicker, it doesn’t mean that suddenly everything is 1000 times faster,” Magouyrk said. Oracle is still working out how to redesign its testing, validation, deployment and release management processes to keep up.
Security work showed a similar pattern. Oracle got early access to Anthropic’s Mythos Preview model, which scans code for vulnerabilities, and found more potential security issues in its first two weeks than in an entire year, Evans said. But she estimated 60% to 70% were false positives, so Oracle built new processes to verify vulnerabilities before engineers start fixing them.
Business Insider noted that other companies are hitting the same cost and workflow limits, citing JPMorgan’s recent decision to set cost limits for employees using Anthropic’s Claude.


WebProNews is an iEntry Publication