A recent study by Economist Impact, commissioned by Databricks, has revealed that while generative AI (GenAI) adoption is widespread among global enterprises, only 22% of organizations believe their IT infrastructure is prepared to support the next wave of AI applications. The report, titled *Unlocking Enterprise AI: Opportunities and Strategies*, surveyed 1,100 technical executives and practitioners across 19 countries and highlighted the complexities of scaling AI across industries.
The findings indicate that 85% of enterprises are currently using GenAI in at least one operational area. Despite this, challenges such as high costs, skill shortages, and concerns over governance and quality are hindering broader adoption. Only 37% of executives and 29% of practitioners regard their GenAI applications as ready for production. Governance issues, noted by 40% of respondents, are among the top barriers, with half of data engineers identifying it as the most time-consuming aspect of their work.
Andy Kofoid, President of Global Field Operations at Databricks, emphasized the need for robust platforms that integrate data, analytics, and governance to help organizations fully realize the potential of AI. “This report showcases why data intelligence is essential and why successful enterprises will adopt a holistic approach that combines data management, governance, and domain-specific expertise,” he stated.
The report also highlighted the potential of emerging AI applications, such as ‘Agentic AI,’ which allows natural language interfaces to streamline complex tasks. Nearly 60% of respondents predict that within three years, natural language will become the primary means for non-technical staff to interact with data. Enterprises are leveraging AI to improve operations across various sectors, including customer service, fraud detection, and healthcare.
Accenture’s Senthil Ramani noted that while AI adoption is a logical starting point for improving workforce productivity, organizations aiming for long-term leadership in AI must focus on growth, customer experience, and risk management. This approach, he stated, will not only enhance efficiency but also open new business opportunities.
The study also shed light on evolving trends in AI technology. Nearly half of data scientists currently use general-purpose large language models (LLMs) that lack integration with proprietary data, leading to quality and governance challenges. However, 58% are now employing retrieval-augmented generation (RAG) to enhance models with contextual enterprise data. Furthermore, 96% of organizations plan to deploy open-source AI models by 2027, reflecting a shift toward flexible, mixed-model systems.
Despite the momentum, organizations face significant hurdles in securing AI talent, with only one in six respondents expressing confidence in their ability to attract the necessary expertise. Investment in both technical and non-technical areas is seen as inadequate by a majority of those surveyed.
Tamzin Booth, Editorial Director of Economist Impact, summarized the findings by emphasizing the importance of integrating proprietary data with AI to develop “data intelligence” and deliver high-quality outputs. According to Booth, enterprises must address governance, performance evaluation, and workforce integration to fully harness AI’s transformative potential.
