A global study into AI adoption—based on a survey of 1,100 C-suite executives and technologists, and interviews from 28 CIOs—reveals a yawning gap between ambition and reality. Firms grapple with antiquated infrastructure, data governance and the delicate balance of human-machine intelligence.
A heady US$1trn in capital expenditure is expected in the coming years to deliver the data centres, chips, energy and infrastructure to support the artificial intelligence (AI) revolution. While this figure is largely driven by the dawn of generative AI (GenAI), the truth is that companies have been developing their AI capabilities for years. Breakthroughs like deep learning and neural networks have allowed specialists in fields from biotech to finance to crunch vast datasets to uncover patterns and deliver actionable intelligence.
The difference now is democratisation and scale. Because GenAI provides an intuitive, natural language interface, the benefits of AI have become accessible to every practitioner. To transition successfully from pilots to enterprise-wide deployment requires a robust infrastructure that can handle the data and computational requirements of AI, a workforce strategy that appropriately calibrates humans versus machines, and an appropriate return on investment (ROI) strategy. Last, but not least, achieving AI excellence necessitates strong governance and careful design of human machine interaction. Decisions and outputs must comply with extensive existing and emerging legislation in areas like data privacy, security and consumer protection.
This Economist Impact report, commissioned by Databricks, combines a global survey of 715 technical executives and 385 data and AI technologists who work across the fields of data engineering, data science and enterprise architecture. It also features insights from interviews with 28 C-suite executives from leading organisations across ten countries, who represent 11 industries.
Key findings include:
-
Executives and practitioners alike believe in the power of AI, but think investment is falling short. From accelerating drug development to extending credit to the financially excluded, companies are finding a diverse array of applications for AI and GenAI. Our survey found that 85% of organisations are actively using GenAI in at least one business function, reaching 97% of companies with revenue over US$10bn. IT teams are the most avid adopters and the legal function is the most reticent. Internal projects are preferred but, by 2027, 99% of executives expect GenAI adoption across internal and external use cases. Seven in ten see AI as crucial to their long-term strategic goals, and only 18% say it is overhyped. Despite the momentum, only one in five believe that their current investment across technical and non-technical domains is sufficient.
-
AI’s short-term benefits include productivity and efficiency, but leaders will use AI to unlock value, revenue and business model innovation in the long term. So far, productivity is one of the most commonly reported impacts of AI. In line with this, functions with high automation potential have been heavily testing GenAI, such as IT (91%), marketing (85%), sales and customer service (83%), and operations (80%), while 82% of data scientists report using AI for coding. Interviewees also report a range of benefits, from cost reduction to improved employee experience and talent attraction. However, executives do take a ‘long’ view of the GenAI paradigm shift, noting that strategic considerations like business model innovation, market positioning, and environment, social and governance standards have been the most important elements when evaluating business cases for AI. Revenue growth has been the least effective metric for justifying investment to date, with experts arguing that AI returns will take time to accrue given the need for experimentation, iteration and digital infrastructure overhauls. Over time, however, the ability to realise revenue will set leaders apart, and financial value will need to be more definitively quantified.
-
Enterprise-wide AI requires an infrastructure refit—most companies do not feel confident in their current architecture. Many enterprises are operating with the technological equivalent of Victorian-era plumbing. Only 22% of organisations say their current architecture can support AI workloads without modifications, and 48% of data engineers spend most of their time resolving data source connections. The problem could worsen as democratisation leads to a proliferation of AI pilots, leading to congestion, complexity, and opacity in data and infrastructure. The prize is worth the work; some of the companies interviewed for this programme reported that they could be truly creative in finding use cases and achieving returns only once they had secured their data foundation. Worries about data security, from silos to fragmented systems, are holding some companies back from more ambitious experimentation.
-
Two-thirds of organisations see significant potential in integrating GenAI models with their own data. As AI models become commoditised, the best performers will mix and match models and tools with their unique assets: proprietary data and know-how. This requires finding the right blend of tools, such as open- and closed-source models, and gaining visibility and mastery over internal datasets. Organisations are exploring a variety of models, and nearly seven out of ten are experimenting with or have fully deployed open-source GenAI, with 96% believing they will do so by 2027. Experts warn that off-the-shelf models are inferior in areas like domain-specific lexicon, and may provide less control and security. However, off-the-shelf models, unconnected to enterprise data, are still common: almost half (45%) of data scientists are using large language models (LLMs) without retrieval augmented generation (RAG).
-
Few enterprises have fully ‘productionised’ GenAI due to cost, skills and governance challenges. Only 37% of executives believe their GenAI applications are production-ready, a figure that falls to just 29% among practitioners. Data scientists cite key constraints including cost (41%), skills (40%), quality (37%) and governance (33%). Only one in six respondents believe that their organisation can secure enough AI talent, half of data engineers say governance takes up more time than anything else, and more than half of enterprise architects (53%) cite challenges with data privacy and security breaches as the biggest risk of AI expansion. Executives agree that in the years ahead, winners will be those that graduate from experiments to production, scaling and monetisation.
- The inevitability of mass AI adoption underscores the critical need to carefully calibrate human and machine intelligence to enable democratised AI ecosystems. Leaders say AI can augment rather than replace employees. Interviewees cite unified systems and ‘centres of excellence’ as critical to balancing governance and enablement, pointing to an irreplaceable role for experience and judgement in overseeing AI outputs. Two-thirds of organisations say they are still experimenting to find the right balance between humans and AI. Many are also investing in self-service tools and AI assistants to empower all employees to become data scientists. Enterprise architects see GenAI as a democratising force, with nearly 60% predicting that within three years natural language will be the primary—if not only—way that non-technical staff interact with complex datasets.