Enterprise AI is generating business insights but not saving money, and the governance gap is widening

Enterprise AI spending is rising across the board, but the returns are landing somewhere most organizations did not plan for. An SAP survey cited by CIO Dive found that companies report AI is helping them surface business insights and improve customer interactions, but it is not delivering the cost savings or time efficiencies that most teams built their business cases around. That disconnect is forcing IT and operations leaders to reassess both their deployment strategies and their vendor relationships.

The platforms absorbing most of that investment are well-established. Microsoft Copilot, embedded across Microsoft 365, is seeing broad enterprise uptake for capabilities including automated document summaries and contextual suggestions during workflows, according to TechRadar AI. Google has extended AI across its Workspace suite to handle document management and workflow automation, while Salesforce's Einstein GPT is being used by enterprise sales and service teams to automate customer interactions and generate personalized responses. Competition among these SaaS incumbents is accelerating adoption timelines for companies that might otherwise move more cautiously.

ROI is real, just not where teams budgeted for it

The SAP findings, as reported by CIO Dive, present a specific challenge for operations and finance leaders who approved AI budgets on the premise of headcount efficiency or process cost reduction. If the measurable gains are instead appearing in insight generation and customer engagement quality, the KPIs used to evaluate those deployments may need to change before the next budget cycle. Teams that cannot demonstrate value against the original metrics risk losing funding even when the technology is genuinely performing.

Agentic AI is compounding the cost-management problem. According to CIO Dive, OpenAI has noted that CIOs need clear visibility into AI demand, spend, and risk to determine whether the technology is generating value as agentic use cases drive consumption up. Agentic systems, which operate with greater autonomy and trigger downstream actions across enterprise environments, can scale usage and cost faster than traditional copilot-style tools, making governance and spending controls a near-term operational priority rather than a future concern.

Cyber AI adoption is outrunning governance frameworks

CIO Dive also reports a sharp increase in AI adoption specifically for cyber defense, and flags a major governance gap that is widening alongside it. Security teams are deploying AI-powered tools faster than policy frameworks can accommodate them, creating exposure that sits at the intersection of IT, legal, and compliance functions. For CIOs and CISOs who report to the same executive team, this is not an abstract risk: it is an audit finding or a breach inquiry waiting to happen.

The White House's launch of a vulnerability clearinghouse, reported by CIO Dive, reflects the broader pressure on enterprises to track AI-fueled surges in security flaws more systematically. The volume of new vulnerabilities associated with AI-generated or AI-assisted code is increasing faster than traditional patch management cycles were designed to handle. Procurement teams evaluating AI coding tools or AI-assisted development platforms should factor vendor disclosure practices and patch cadence into their selection criteria.

A trust deficit that procurement cannot ignore

Nearly half of AI users distrust the companies behind the tools they rely on daily, according to a report from fintech firm Tether cited by CIO Dive. That figure is striking because it sits alongside data showing chatbot and AI assistant usage is widespread and still growing. The implication for enterprise buyers is direct: employees are using AI tools they do not fully trust, which creates shadow IT behavior, inconsistent adoption, and data handling risks that IT policy has to address explicitly.

TechRadar AI notes that regulatory compliance is one of the reasons enterprises are turning to AI platforms in the first place. But if underlying user trust in AI vendors is low, compliance use cases that depend on employee engagement, such as automated reporting, policy acknowledgment workflows, or AI-assisted audit trails, may underperform. Vendor transparency on data handling, model behavior, and audit logging is no longer a differentiator; it is a baseline procurement requirement.

What this means for your team

• Audit your AI ROI metrics now: if your business case was built on cost savings or time reduction, compare it against actual outcomes. The SAP survey data suggests insight generation and customer interaction quality may be where value is actually accruing, and your reporting framework should reflect that.
• Build AI governance ahead of the next deployment, not after. CIO Dive's coverage of the cyber defense governance gap is a direct signal that security and policy teams need to be in the room before the next AI tool goes into production, not during the incident review.
• Add trust and transparency criteria to your AI vendor scorecard. With nearly half of users distrusting AI companies, procurement teams should require vendors to document data handling practices, model governance policies, and audit log access as conditions of contract.
• Establish spend visibility for agentic AI specifically. Agentic systems scale consumption differently than assistant-style tools; CIOs should set usage thresholds and cost alerts before agentic deployments go broad.