Domino Data Lab Report: 32% Partial Governance for Agentic AI Erodes ROI in Regulated Firms

Agentic AI governance is the set of formalized policies, monitoring mechanisms, and compliance controls that organizations apply to autonomous AI agents capable of independent planning and execution.

The BARC guest post hosted by Domino Data Lab identifies a clear pattern where agentic AI projects in regulated industries encounter ROI shortfalls when governance remains incomplete. Organizations that implement only partial structures for oversight of autonomous agents find that expected efficiency gains do not materialize into measurable returns.

This gap occurs because many enterprises prioritize rapid deployment of agentic systems to address operational challenges without first establishing complete governance layers. The result is a structural mismatch that affects performance tracking and compliance verification in sectors subject to strict regulatory scrutiny.

What Background Context Surrounds Agentic AI Adoption in Enterprises?

Agentic AI systems differ from earlier generative models by incorporating planning loops and tool use that allow them to pursue goals with reduced human intervention. Enterprises in finance, healthcare, and manufacturing have begun integrating these systems to automate complex workflows that previously required multiple teams.

The acceleration of adoption stems from competitive pressures and the promise of cost reduction. However, the same autonomy that delivers operational speed also introduces new categories of risk around decision traceability and regulatory alignment that partial governance cannot fully address.

Regulated organizations operate under frameworks that demand auditable processes at every stage. When agentic AI operates outside fully formalized controls, the ability to demonstrate compliance during audits diminishes, creating exposure that directly affects financial projections tied to these investments.

What Technical Specifics Define Partial Governance in Agentic AI?

Partial governance typically includes basic policy documentation and some monitoring tools but omits real-time decision logging, automated compliance checks, and comprehensive risk scoring for autonomous actions. This incomplete setup leaves gaps in visibility into how agents select and execute tasks.

Full governance would require integration of agent activity with existing enterprise risk systems, continuous validation against regulatory rules, and mechanisms to halt or redirect agent behavior when thresholds are breached. The absence of these elements in 32% of cases prevents organizations from confirming that agent outputs align with business and legal requirements.

Technical debt accumulates when organizations attempt to retrofit governance after initial deployment. Retrofitting proves more costly than building controls concurrently with agent development, contributing to the ROI erosion documented in the BARC analysis.

What Market and Stakeholder Implications Arise from These Findings?

Stakeholders including CFOs and compliance officers must recalculate expected returns on agentic AI when governance coverage remains partial. Budget allocations that assumed full operational efficiency now require adjustment to account for remediation expenses and potential regulatory penalties.

The broader market may experience slower uptake of agentic AI as procurement teams demand evidence of governance maturity before approving large-scale rollouts. Vendors that provide integrated governance platforms alongside agent capabilities stand to benefit from this shift in buyer priorities.

Regulated firms in particular face asymmetric risk because their revenue models depend on maintaining licenses and avoiding enforcement actions. Partial governance increases the probability that agent-driven decisions will trigger reviews that delay projects and consume resources originally allocated to innovation.

What Expert Reactions Highlight the Governance Disconnect?

The BARC guest post serves as a direct warning to enterprise leaders that current deployment velocities exceed governance development rates. Analysts emphasize that the 32% partial governance figure represents a systemic issue rather than isolated implementation errors.

agentic AI deployments outpace governance (32% partially formalized), directly causing ROI breakdowns in regulated enterprises.BARC guest post

What Steps Should Enterprises Take Next?

Enterprise teams need to conduct maturity assessments that map existing controls against the full requirements for agentic systems. This assessment identifies which components of governance are missing and prioritizes remediation based on regulatory exposure.

Investment in platforms that natively support governance features for autonomous agents can reduce the time required to reach full coverage. Collaboration between data science, legal, and risk management functions ensures that technical implementations satisfy compliance standards from the outset.

Ongoing monitoring programs that track both agent performance and governance adherence provide the feedback loop necessary to maintain alignment as agent capabilities evolve. Without such programs, organizations risk repeating the partial governance pattern at larger scales.

Governance maturity levels and associated ROI effects drawn from BARC analysis hosted by Domino Data Lab.

Governance Level	Reported Share	Primary ROI Impact
Full governance	Not specified in report	Sustained returns with compliance assurance
Partial governance	32%	Direct erosion of projected returns in regulated sectors
Minimal or absent governance	Not specified in report	Elevated risk of regulatory intervention and project failure

1. Conduct a formal governance maturity assessment across all agentic AI projects.
2. Map regulatory requirements to specific agent decision points and logging needs.
3. Deploy integrated monitoring tools that cover autonomous planning and execution.
4. Establish cross-functional review cycles involving compliance and technical teams.
5. Implement continuous audit mechanisms tied to agent activity logs.
6. Update governance policies quarterly to reflect changes in agent capabilities and regulations.