Generative artificial intelligence fascinates with its innovative potential, but companies face significant obstacles. *97% of organizations struggle to clearly present the return on investment.* Technological ambitions are accompanied by a complex reality marked by challenges in demonstrating value. *The fight to prove the impact on commercial performance* highlights excessive expectations in the face of often incomplete results. A persistent momentum, despite a contrasting picture, underscores the urgency of redirecting efforts towards solid foundations.
The challenge of demonstrating the value of generative AI
A majority of companies engaged in implementing generative artificial intelligence encounter major obstacles in proving its impact on commercial performance. According to the CDO Insights 2025 report published by Informatica and Wakefield Research, an alarming 97% of data leaders admit to struggling to quantify the return on investment of their initiatives.
Rising investments despite uncertainty
Despite the difficulties encountered, a paradox arises. Nearly 87% of organizations that have adopted or are considering the adoption of generative AI plan to increase their investments in this area by 2025. A significant proportion of these companies, around a quarter, expect a substantial increase in these expenditures.
The United States stands out in this dynamic, with 93% of respondents reporting an increase in investments, compared to 82% in Europe and 86% in Asia-Pacific. This desire to invest does not temper the uncertainty of their results. The inability to prove the business value of these technologies disrupts initial enthusiasm.
Obstacles to demonstrating added value
The most frequently cited barriers by companies include challenges related to cybersecurity and compliance (46%), concerns about responsible use of AI (45%), as well as doubts regarding the reliability of the results obtained (43%). Trust in the quality of data is also lacking for 38% of the surveyed companies.
A gap between expectations and reality
The report highlights the growing gap between the expectations of data leaders and those of executives. Nearly 92% of data leaders believe that executives want results faster than what the technology can offer. In the United States, this figure reaches 97%, highlighting immense pressure on teams to quickly produce tangible results.
Corporate priorities: data quality and training
Faced with the challenges encountered, companies are now redirecting their strategies. A majority of about 86% of organizations plan to increase their investments in data management by 2025. The reliability and preparedness of data are emerging as priorities to maximize the potential of generative AI.
Organizational challenges include an urgent need to develop data literacy within teams. Training remains crucial, as nearly all surveyed companies report inappropriate uses of generative AI, causing compliance issues and the exploitation of incomplete data.
Future perspectives
The transition from pilot phase to production appears complex. About 67% of organizations have failed to move half of their pilot projects into production. Technical obstacles, such as poor quality data and a lack of maturity in the technologies used, present additional challenges.
The study also reveals that without a clear definition of KPI at the launch of projects and a shared understanding of what “value” entails, demonstrating the benefits of AI remains hindered. The desire to accelerate the use of generative AI runs into sometimes disenchanted realities, raising questions about the future of these investments.
Frequently asked questions about generative AI and its impact on commercial performance
Why do companies struggle to prove the value of generative AI?
Companies face difficulties in demonstrating the value of generative AI due to obstacles such as a lack of reliable data, unrealistic expectations from executives, and issues related to cybersecurity and compliance. These challenges make it hard to evaluate returns on investment.
What are the main reasons why 97% of companies find it hard to value generative AI?
Reasons include the difficulty of establishing performance indicators (KPIs), the experimental nature of projects, and technical obstacles such as data quality and a lack of AI skills within teams.
How do investment ambitions in generative AI compare with the achieved results?
Despite results being difficult to prove, 87% of companies plan to increase their investments in generative AI. This shows a gap between enthusiasm for technology and the reality of measurable benefits.
What role do executive expectations play in the adoption of generative AI?
Unrealistic expectations from executives, who hope for quick results, can lead to a proliferation of pilot projects without solving previous issues, thereby compromising the credibility and effectiveness of initiatives.
What solutions are companies considering to overcome these challenges?
Organizations are looking to enhance data quality and team training, with 86% of them planning to increase their spending on data management to ensure better preparation and reliability for generative AI projects.
How does training impact the adoption of generative AI?
Proper training is essential to avoid inappropriate uses of generative AI. On average, it takes 11 months for employees to reach a sufficient skill level to use these tools effectively and responsibly.
