The front-runners’ guide to scaling AI (Accenture, 2025)
TL;DR
An Accenture Research report published on 6 May 2025. It surveyed 2,000 executives at 1,998 companies with revenues above $1 billion, in June and July 2024 (p. 3). It sorts the companies into four groups by their maturity on ten data and AI capabilities.
- 8% are “front-runners”: “Only 8% of companies—‘front-runners’—are scaling AI at an enterprise level, embedding the technology into core business strategy” (landing page; Figure 2, p. 9).
- They are a subset of 15% “AI reinvention-ready” companies. The other 7% are “fast-followers”, which have the capabilities but “have yet to scale any strategic bets” (p. 9). The rest are “progressing with AI” (43%) and “experimenting with AI” (42%) (p. 8).
- Even front-runners have scaled only a third. On average they have scaled “34% of the strategic bets … most relevant to their industry”. Another 40% are in early stages (p. 11).
- A broader figure sits underneath. “34% of surveyed companies have scaled at least one strategic bet” (p. 5). Those companies are “nearly three times more likely to have their return on investment (ROI) from gen AI surpass their forecasts” (p. 6).
- Financial gap. Front-runners with revenues above $10 billion grew revenue 7 percentage points faster in 2023 than companies still experimenting. Their ROIC was 4 points higher and their 2019–24 TSR 6 points higher than the other three groups (pp. 20, 36). The authors add that “further research is needed to establish a causal link” (p. 36).
Key data
| Group | Share | Definition | Page |
|---|---|---|---|
| Front-runners | 8% | AI reinvention-ready and “have already scaled multiple strategic bets” | 9 |
| Fast-followers | 7% | AI reinvention-ready, no strategic bet fully scaled yet | 9 |
| Progressing with AI | 43% | “intermediate levels of maturity” | 8, 9 |
| Experimenting with AI | 42% | “comparatively low levels of maturity” | 8, 9 |
| Figure | Definition | Page |
|---|---|---|
| 15% | AI reinvention-ready (front-runners + fast-followers) | 8 |
| 34% | companies that have scaled at least one strategic bet | 5 |
| 34% / 40% | front-runners’ strategic bets scaled / in early scaling | 11 |
| 12% / 2% | share of front-runners in life sciences / retail (average 8%) | 15 |
| 70% | companies that see the need for a strong data foundation to scale AI | 4 |
| one-third | companies already using AI agents to strengthen innovation | 4 |
| 13% / 12% / 11% / 11% | expected productivity / revenue growth / customer experience / cost improvement within 18 months of scaling gen AI | 6 |
| 51% vs 45% | share of tech budget on cloud and AI, companies that scaled a strategic bet vs those that did not | 5 |
| ±2.2 points | stated margin of error at 95% confidence | 5 |
Method and definitions
- Sample (p. 3): “2,000 C-suite and data-science executives, who lead 1,998 of the world’s largest companies (revenues greater than $1 billion)“. The companies are headquartered in 15 countries and work in nine industries: banking, insurance, energy, consumer goods and services, life sciences, utilities, retail, public services, and communications and media. Health-care providers are not among them. The survey was “fielded from June to July 2024.”
- Sample filter (p. 35): “Our sample is comprised of companies that have, at the very least, a basic AI strategy in place.” Companies without one are excluded, so the shares are shares of AI-active large firms.
- Strategic bets (p. 5): “significant, long-term investments in gen AI that focus on the core of a company’s value chain.” Accenture experts and outside AI experts chose 105 of them, “just over 11 per industry”. Respondents rated adoption of each bet for their industry from “no adoption” to “full scaling.” Table stakes are the broad, foundational investments.
- Grouping (pp. 35–36): companies were clustered on ten capabilities. Five are “foundational” (data and AI strategy, AI platform, data management and governance, talent, responsible AI). Five are “new essentials” for gen AI (LLMOps, new data management, data sources, foundation-model practice, talent practice). The method is hierarchical clustering with Ward’s linkage and Gower’s distance. A latent-class check agreed on 85% of companies.
- “Scaling AI” (p. 3): “expanding AI implementation across an enterprise to achieve broader, more impactful outcomes,” including integration into processes, widespread adoption and improved key metrics.
Scope and reliability
- Older data than the date suggests. Published May 2025, fielded mid-2024. Quoting the 8% as a 2025 measurement overstates its recency.
- Self-report by a seller. Executives rate their own capabilities and adoption. Accenture sells AI and data transformation services, and the report ends with its “five imperatives” (pp. 21–31). It is not peer-reviewed.
- Filtered sample. Large firms with at least a basic AI strategy only (p. 35). The 8% is a share of those, not of all companies.
- Correlation. The financial gaps are correlations “after controlling for region and industry effects” (p. 36). The 7-point revenue-growth gap holds for front-runners above $10 billion in revenue, against experimenters only.
- Unit. One respondent per company in almost all cases (2,000 executives, 1,998 companies; p. 5).
Dynamic-capabilities reading
digital-transforming/improving-digital-maturity: the grouping rests on maturity in ten capabilities, including talent maturity. Front-runners have “a talent maturity that is four times higher” than experimenters (landing page).digital-seizing/balancing-digital-portfolios: the report separates “table stakes” from “strategic bets” and measures how far each bet has scaled. That is portfolio balancing and scaling up.contextual/internal-enablers: executive support. When strategic bets are sponsored by the CEO or board, “these investments are 2.4x more likely to exceed their projected ROI than investments that don’t receive such sponsorship” (p. 22). 74% of the companies surveyed have a Chief AI Officer or similar C-suite role (p. 22).
How it connects
- enterprise-ai-adoption: a capability-based scaling figure (8%) for large, AI-active firms.
- ai-maturity-measurement-comparison: a strict-definition instrument fielded in 2024. Its broader figures (15% reinvention-ready; 34% with one bet scaled) come from the same survey.
- BCG’s Widening AI Value Gap (2025): a similar design, capability scores clustered into four groups, with a 5% top group.
- Dutt et al. (Bain and OpenAI) and micro-productivity-trap: a few core-process bets against broad, shallow adoption.
- Accenture’s TQ Tech Talk on AI harnesses: the other Accenture source in the wiki.
- responsible-ai: one of the five imperatives (“Close the gap on responsible AI”, pp. 27–28).
What was actually ingested
The full 39-page PDF linked from the landing page, converted with pdftotext -layout, and the landing page itself (curl). Group shares were read from the text layer of Figures 2 and 5 (pp. 9, 15). Appendix 1, the table of 105 strategic bets by industry, was skimmed, not transcribed.
Linked entities and concepts
- Accenture — publisher (Accenture Research).
- enterprise-ai-adoption, micro-productivity-trap, responsible-ai: concepts.
- ai-maturity-measurement-comparison: thread.
Dangling (single-source mention, deferred): Senthil Ramani (Global Lead for Data & AI, Accenture), Lan Guan (Chief AI Officer, Accenture), Philippe Roussiere (Global Lead for Innovation and AI, Accenture Research).