While Gen AI adoption surges, implementation at scale remains challenging
Generative AI is already making a significant impact across functions and industries, but many organisations are struggling to keep pace with its potential and the rapid rate of change. This is according to a survey conducted by SparkOptimus among more than 50 companies across various sectors.
For the past several years, there has been a race to adopt AI and its little sister Gen AI, with companies enticed by the impressive additional value that these tools can bring. Indeed, when Gen AI tools are well integrated into a company’s business model, it can significantly boost efficiency, effectiveness, insights, and much more.
This year’s edition of the annual ‘Gen AI Benchmark’ by SparkOptimus found that more companies are running pilots in 2025, they are setting clearer goals, scaling their first use cases, and Agentic AI is becoming more promising. As the technology rapidly matures and more companies leverage these types of tools, it is increasingly a major part of staying competitive.
Nearly every respondent surveyed said that they use Gen AI (92%), compared to 60% last year. This shows how the technology is becoming unavoidable and highlights the need to integrate it in a smart way into organizations. 59% said they use it a few times a month while 25% said they use it multiple times a week.
Three quarters of companies (75%) have a Gen AI strategy in place. That is compared with only 40% from last year’s survey. Notably, a growing number of companies now also have a standalone strategy (just 10% last year).
Formal data strategy
Having a good data strategy in place often leads to more tested use cases. The percentage of companies that report having a formal data strategy in place is up from last year. The number of use cases that go from idea to pilot for companies with a formal data strategy increased by 19 percentage points.
Despite that, 45% of companies still use unstructured data, which can lead to issues with results. The report also found 60% of organizations lack the tools to keep unstructured data to-to-date, which also tends to lead to headaches.
Challenges along the way
Despite the significant progress booked in ambitions and pilots, the report at the same time emphasizes that integrating AI into daily processes and ways of working has not been an easy task for many.
The survey found that many employees lack skills in using Gen AI effectively and that only 20% are upskilled. Upskilling remains a challenge for many companies that have not invested sufficiently in teaching employees how to best use AI tools.
With 60% of respondents self-reporting their Gen AI skills as ‘insufficient’ in many cases, employees need more guidance. The lack of upskilling in many organizations is likely holding back further adoptions and the potential impact that Gen AI can have. With the arrival of the AI literacy pillar of the EU AI Act, organizations are expected to grow their focus on this matter.
Elsewhere, only 27% of companies back their Gen AI strategy up with quantified ambitions. This lack of solid ambitions makes it harder to measure progress towards the goals and specific outcomes that companies hope to achieve.
Another interesting result of the survey is that many leaders feel that improvements can be made to the way Gen AI is piloted. Many pilots go on for too long and end up not leading to viable use cases, with around half of pilots running for more than three months. And when pilots are deemed successful, the transition to real life use cases or scaling often stalls.
“While progress is undeniable, the rollout of Gen AI at scale remains difficult,” said Matti can Engelen, associate partner at SparkOptimus. “Many organizations face common roadblocks like fragmented ownership, lack of integration into core business processes, and unclear value metrics. As a result, efforts often remain siloed, with real impact still constrained to isolated teams or use cases.”
