From pilots to enterprise impact, one year on
Generative AI has sped from proof-of-concept to enterprise agenda, but only a small cohort is converting pilots into measurable value at scale. Accenture’s research across 3,450 executives and 2,000 projects confirms five imperatives—lead with value, build a secure digital core, close the responsible AI gap, drive continuous reinvention, and reinvent talent—as the playbook for advantage. The leaders pair agentic architectures, proprietary data, and people-centric change with CEO sponsorship to compound gains in 2025.
Points clés
- Accenture’s analysis draws on over 2,000 gen AI projects and a Pulse of Change survey of 3,450 C‑suite leaders across 22 industries and 20 countries.
- Only 36% of surveyed executives have scaled gen AI solutions and just 13% report significant enterprise-level value, while budgets are spent 3x more on technology than on people.
- Seventy‑three percent of gen AI investment is concentrated in IT, customer service, and marketing; 34% of organizations have scaled at least one industry‑tailored core‑process solution and are 3x more likely to exceed ROI expectations.
- Eighty‑three percent of executives say gen AI’s potential now exceeds expectations; organizations already delivering enterprise value are 4.5x more likely to invest in agentic architecture and 6x more likely to increase gen AI spending in 2025.
- Acting across all five imperatives makes organizations 2.5x more likely to achieve enterprise‑level results; CEO sponsorship correlates with 2.5x higher ROI, and leaders who succeed are 6x more likely to deeply understand gen AI.
- Organizations creating enterprise value are 2.9x more likely to have a comprehensive data strategy and 2.7x more likely to have responsible AI governance; advanced change capabilities raise transformation success odds by 2.1x.
- Ecolab is reinventing its Lead‑to‑Cash value chain using utility, super, and orchestrator AI agents, a new AI operating model, and capability building to unlock cross‑functional value.
- Accenture’s Marketing + Communications deployed AI Refinery with 14 specialized agents for 954 marketers, cutting internal communications 60% and external content 50%, lifting brand value 25%, reducing manual work ~30%, and improving speed‑to‑market 25–55%.
- Sempra’s SoCalGas and San Diego Gas & Electric modernized a secure cloud digital core with data mesh and responsible AI governance, cutting analysis time by up to ~90% and improving asset reliability and customer service.
- A large electronics retailer improved intent identification (chat +14%, IVR +35%), intent containment (chat +70%+, IVR +100%+), and agent productivity (30M summaries; −38s chat and −30s voice handle time), while a leading bank cut legal reviews from three days to one and advanced eight enterprise-grade pilots.
À retenir
Start with CEO sponsorship, not a thousand pilots—your odds of real ROI go up faster than your coffee bill. Rebalance spend toward people (yes, the ones actually using the tools) and pick one mega‑process like Lead to Cash to reimagine with agents plus humans. Lock in a data strategy, bake responsible AI into the foundation, and build a modern, modular digital core—because duct tape is not an architecture. Then measure everything, communicate often, and keep iterating; continuous reinvention beats “set‑and‑forget” every time, even if your calendar disagrees.
Sources
Quiz sur le document: 10 questions
