Accenture: How Enterprises Can Close the AI Scaling Gap

Accenture reports that while most large enterprises are experimenting with AI, only a small fraction are successfully scaling AI initiatives across their operations to drive meaningful business transformation.

The research, which surveyed 2,000 C-suite and data science executives from 1,998 companies with revenues exceeding US$1bn, found that just 8% of organisations qualify as “front-runners” – companies that have successfully scaled multiple strategic AI implementations. 

These front-runners have achieved an average of 34% scaling across their industry-specific AI strategic bets.

“Today, our clients need more value faster and Accenture is their reinvention partner of choice,” says Julie Sweet, Chair and CEO, Accenture. 

“We are writing the playbook for how to be the most AI-enabled, client-focused professional services company in the world.”

Authors of the report: Senthil Ramani, Global Lead for Data & AI at Accenture, Lan Guan Chief AI Officer (CAIO) at Accenture and Philippe Roussiere, Global Lead for Innovation and AI, at Accenture Research, say: “For businesses, securing a sustained advantage over competitors was long the Holy Grail – a coveted, yet elusive prize. 

“Today, however, Gen AI and other forms of AI have flipped the script, bringing the previously unattainable within reach.”

Accenture’s three company categories

The study, conducted between June and July 2024, examines companies across nine industries including banking, insurance, energy, retail and life sciences. 

Accenture defines “scaling AI” as expanding AI implementation across an enterprise to achieve broader outcomes, including integrating AI into business processes, ensuring widespread adoption – and improving key performance metrics.

The consultancy classified organisations into three categories based on their AI maturity, described as:

• 42% are “experimenting with AI” 
• 43% are “progressing with AI”
• 15% achieve “AI reinvention-ready” status
• Within the 15%, 8% classified as front-runners have scaled multiple strategic bets

Strategic bets, as defined by Accenture, are significant long-term investments in Gen AI that focus on core value chain activities such as underwriting for insurers or R&D for pharmaceutical companies. 

These differ from “table stakes” investments, which are foundational AI implementations like customer support chatbots that offer incremental rather than transformative value.

The authors note that front-runners demonstrate superior capabilities in what the firm terms “new data and AI essential capabilities for Gen AI”.

These include large language model operations (LLMOps) maturity, advanced data management practices and specialised talent development.

The research also identifies 105 strategic bets across the nine industries studied. 

Companies in utilities focus on workforce operations optimisation and generation forecasting, whilst banking institutions prioritise fraud management and cards and payments systems. 

Meanwhile, life sciences companies concentrate on accelerating time to market and clinical trials.

Financial performance gap widens between leaders

Front-runners significantly outperform other categories in financial metrics. 

In 2023, these companies achieved revenue growth 7% higher than firms merely experimenting with AI. 

Their return on invested capital exceeded other groups by 4%, whilst total shareholder returns were 6% higher during the 2019-2024 period.

Accenture says that companies that have scaled at least one strategic bet are nearly three times more likely to exceed their projected return on investment from Gen AI compared to those that have not scaled any strategic initiatives. 

Front-runners additionally report expectations of 13% productivity increases, 12% revenue growth, 11% customer experience improvements and 11% cost reductions within 18 months of enterprise-wide AI deployment.

As a result, the authors emphasise the importance of agentic architecture – networks of AI agents that orchestrate business workflows rather than simply automating routine tasks.

One-third of surveyed companies already deploy AI agents to strengthen innovation capabilities and 70% of companies acknowledge the need for strong data foundations when scaling AI.

“The future belongs to those who can scale AI not just in pilots – but across every domain of the supply chain,” says Kristalyn Warren Mumaw, AI Evangelist at Accenture.

However, low data readiness remains a primary obstacle, particularly regarding unstructured data utilisation.

Outdated IT systems and insufficient worker access to AI tools and training also present significant barriers.

The strategic investment patterns emerging

Accenture says that front-runners allocate 51% of their technology budgets to cloud and AI initiatives, compared to 45% among companies that have not scaled strategic bets. 

These leading organisations demonstrate superior CEO and board sponsorship, with 19% having strong executive backing compared to 5% of “fast-followers” – AI reinvention-ready companies that have not yet scaled strategic bets.

Therefore, the research highlights the importance of centralised operating models. 

While 57% of front-runners employ centres of excellence for AI strategy and deployment, only 16% of fast-followers have implemented similar structures.

Data utilisation patterns differ markedly between groups. 

Front-runners more frequently leverage zero-party data (44% versus 4% for fast-followers), second-party data (30% versus 7%), third-party data (25% versus 8%) and synthetic data (35% versus 6%). 

They also demonstrate advanced capabilities in retrieval-augmented generation (RAG) to enhance large language models, with 17% of front-runners using this technique compared to 1% of fast-followers.

The industry leaders emerging

Life sciences companies show the highest concentration of front-runners at 12% of surveyed firms in that sector, followed by insurance at 10%. 

Life sciences companies prioritise accelerating time to market (16%), accelerating time to clinic (14%) and maximising medicine value propositions (13%).

Yet retail companies lag significantly with only 2% achieving front-runner status.

In insurance, fraud detection represents the most scaled strategic bet at 23% of companies, followed by call assistance at 13% and claims intake at 12%. 

Meanwhile, banking institutions focus primarily on fraud management and cards and payments, each scaled by 29% of firms. 

Five imperatives defining the success path

Accenture’s analysis identifies five imperatives that enable successful strategic bet scaling. 

These begin with “Lead with value”, requiring proactive CEO and board engagement in AI investments with clear value targets and disciplined priorities.

The second imperative, “Reinvent talent and ways of working”, involves broad AI upskilling, dynamic workforce models and human-agent collaboration. 

Accenture believes that organisations must recruit and retain specialists including AI strategists, AI architects and computational scientists whilst building university partnerships for talent pipeline development.

“Build an AI-enabled, secure digital core” is the third imperative according to the report authors, requiring modernised data ecosystems, embedded AI models and agentic architecture aligned with business needs. 

Companies should additionally develop “cognitive digital brains” – AI-powered intelligence hubs for enterprise decision-making that process data streams in real-time.

The fourth imperative, “Close the gap on responsible AI”, extends beyond compliance to value creation through strengthened customer trust and improved product quality. 

AI-driven incidents increased 32% in 2023 according to the AI Incident Database, making responsible AI practices essential for sustainable scaling.

“Drive continuous reinvention” completes the framework, recognising that enterprise transformation requires ongoing adaptation rather than one-time implementation.

This involves integrating change into organisational culture whilst maintaining financial discipline through ROI monitoring and resource reallocation.

Implementation challenges persist across categories

Despite varying AI maturity levels, Accenture finds that companies face similar scaling challenges with different intensities. 

Building and maintaining multi-disciplinary teams is the greatest challenge for both front-runners and experimenting companies, whilst constructing end-to-end data foundations poses the primary obstacle for progressing organisations.

Foundation model customisation, demonstrating concrete ROI and managing security and privacy risks complete the top five challenges identified across all company categories. 

Whereas front-runners show four times greater likelihood than fast-followers to prioritise cultural adaptation and companies with highly developed change capabilities more than double their chances of successful enterprise reinvention.

The research methodology employs hierarchical clustering using Gower’s distance to identify distinct company groups based on 10 critical capabilities. 

These include foundational elements such as data and AI strategy, platform maturity and talent development, alongside new essentials including LLMOps maturity and foundation model practices.

Accenture validated its clustering approach through comparison with latent class analysis, achieving 85% overall concordance and 95% agreement for AI reinvention-ready companies. 

This high alignment reinforces the robustness of the classification methodology and confirms that front-runners exhibit distinct, identifiable characteristics.

Looking ahead, the consultancy emphasises that AI has evolved beyond efficiency improvement to become a force for enterprise reinvention. 

Companies seeking to join the front-runner category must prioritise the five imperatives whilst developing the foundational and new essential capabilities required for sustained AI scaling success.

The authors conclude: “When used to its full potential, AI is now something far greater: an unstoppable force for enterprise reinvention, allowing companies to grow faster and innovate better than rivals.”