Dive Brief:
- Siloed AI systems are restricting AI gains and contributing to sprawl as only 8.2% of organizations have adopted a single shared AI deployment platform, according to FICO’s State of Responsible AI 2026 report. The report is based on a survey of more than 1,000 senior technology, data and risk executives conducted by Corinium on behalf of FICO.
- Just one in five organizations are actively working toward a unified AI deployment platform while 71.2% of organizations have partially adopted single platforms, the report found. Organizations are struggling to agree on one standard for AI development and deployment across different business units, said Scott Zoldi, chief analytics officer at FICO and a contributor to the report.
- “Once that shared standard is in place, that’s the flywheel,” Zoldi told CIO Dive. “At that point you no longer have governance teams stopping anywhere from 50% to 90% of all AI deployment because it’s not sound, it’s not monitored.”
Dive Insight:
Establishing an AI standard — a crucial underpinning for a single shared platform — is the first step toward improving ROI.
The standard should outline key AI development and deployment components such as the algorithms an organization plans to use and how teams will conduct explainability and monitoring, Zoldi said. In addition, businesses should describe how regulatory requirements will be met.
A unified platform then enables the organization’s AI standard to be realized, he added.
“That’s when we see the ROI developing,” Zoldi said.
Along with advocating for and aligning leadership around an AI standard to enable a unified deployment platform, CIOs should also be frugal and responsible with technology purchases, Zoldi said.
CIOs can take a more business value-focused role and help convey the importance of coordination and efficient infrastructure in leading to value for the company rather than adding more tools and creating additional sprawl, he said.
“Very often, that sprawl means shadow projects that are being done, things that are left on that raise your costs along with hyperscaler token use that goes to no business value whatsoever,” Zoldi said.
Only 5% of organizations surveyed for the FICO report said AI significantly exceeded their ROI expectations. More than 37% said AI met their ROI expectations and roughly 43% said the technology somewhat exceeded their expectations.
Security concerns are cited by around 40% of organizations as a barrier to achieving returns on agentic AI, the report found. Another roughly 38% of organizations cited data integration and quality challenges as the biggest barrier.
As enterprises face agentic AI complexity, vendors are responding. Salesforce and Databricks launched AI agent governance features earlier this year. Meanwhile, AWS debuted its Agent Registry platform to help build and govern agents.
