⚡ Quick Summary
The ROI of AI 2025, published by Google Cloud and based on a global survey of 3,466 business leaders, shifts the AI conversation from experimentation to value realization. It introduces a new AI maturity phase—the “agentic era”—where AI agents act autonomously within guardrails to deliver measurable business outcomes. The report documents strong ROI in productivity, customer experience, business growth, marketing, and security, especially for early adopters. With 88% of agentic AI leaders already seeing returns, the report offers a blueprint for scaling adoption, emphasizing the need for executive sponsorship, data governance, and cross-functional integration.
🧩 What’s Covered
The report is structured around four key sections:
- The Agentic Shift (pp. 7–20)
- Defines AI agents as autonomous systems combining LLMs with access to tools and APIs.
- Describes three maturity levels: simple tasks, single-agent apps, and multi-agent workflows .
- Presents industry-wide and global adoption rates: 52% of gen AI users also use agents; 39% have >10 agents deployed .
- Highlights the advantage of early adopters who dedicate 50%+ of their AI budget to agents and see higher ROI across use cases.
- 5 Proven ROI Areas (pp. 21–37)
- Productivity: 70% of leaders report productivity gains; 39% saw ROI from individual productivity tools (email, docs, meetings) .
- Customer Experience: 63% reported improvements; 37% saw ROI in customer-facing applications .
- Business Growth: 56% of firms saw revenue gains, with most estimating a 6–10% boost.
- Marketing: 55% cited impact; gen AI accelerates personalization, journey design, and campaign efficiency.
- Security: 49% report improvements; gen AI enhances threat detection, response speed, and ticket volume reduction .
- Investment Trends (pp. 38–45)
- Reports a sharp rise in AI budgets: 77% increased spending due to falling tech costs.
- 58% fund AI through net new budget lines; 48% reallocate non-AI budgets .
- Top investment drivers: data quality, change management, upskilling, and risk governance.
- Executive sponsorship is critical: 78% of companies with C-suite support report ROI, vs. 43% without .
- Next Steps (pp. 46–47)
- Offers a practical ROI checklist focused on data governance, budget alignment, C-level backing, and talent development.
- Recommends starting with repeatable, high-ROI tasks and embedding human oversight.
💡 Why it matters?
This report marks a clear pivot in the enterprise AI narrative—from exploration to impact. Its focus on agentic AI reflects a major operational transformation: AI is no longer just augmenting decision-making—it’s executing it. For organizations still framing their AI strategies, the benchmarks and investment guidance offered here are invaluable. The emphasis on C-suite sponsorship and cross-departmental ROI reinforces a governance-first mindset, aligning well with the EU AI Act, ISO 42001, and other emerging frameworks. It also validates the business case for responsible AI deployment with hard data on productivity and security gains.
❓ What’s Missing
- Ethical & regulatory context: There is minimal reference to legal, ethical, or safety concerns beyond data privacy.
- Agent governance models: No discussion on how agents should be monitored, overridden, or audited.
- Deployment barriers: The report underplays integration, vendor lock-in, or workforce resistance.
- SME insights: All data comes from $10M+ enterprises; smaller firms are not represented.
- ROI measurement methods: The report reports on ROI but doesn’t explain how it’s calculated or benchmarked internally.
👥 Best For
- AI program leaders scaling from pilot to production
- CTOs and CIOs investing in agentic infrastructure
- Digital transformation and operations teams
- Governance professionals building business cases
- Investors analyzing enterprise AI maturity
- Policy analysts tracking adoption trends
📄 Source Details
- Title: The ROI of AI 2025
- Authors: Google Cloud & National Research Group
- Date: July 2025
- Methodology: Survey of 3,466 senior leaders in global enterprises ($10M+ revenue); fieldwork April–June 2025
- Industries: Financial services, retail, telecom, healthcare, media, manufacturing, public sector
- Link: Not public, internal PDF review
- Pages: 48
📝 Thanks to
Google Cloud’s Oliver Parker (VP, Global Gen AI GTM) and the National Research Group team for producing this longitudinal insight into the business impact of AI agents.
