Global consulting firm Valiantys, in partnership with Atlassian, released new Omdia research on 15 October that starkly illustrates the chasm between AI adoption and actual readiness ([1]). The report, titled Why Enterprises Won Cloud Adoption but Risk Losing AI, surveyed 400 technology leaders worldwide on the barriers to scaling AI. It found that 98% of organisations have at least one AI initiative underway, but only 3% consider themselves fully ready for AI ([2]). Moreover, just 28% of these leaders say even half of their technology transformation projects deliver the expected outcomes ([3]), indicating that many AI efforts are struggling to translate into tangible business value.
The Omdia study pinpoints fundamental data issues as a major cause of this readiness gap. In fact, 44% of respondents identified data quality and accessibility as their single biggest transformation barrier ([4]), far outpacing concerns like legacy systems (cited by 39%). Even after accelerating cloud adoption, with most firms pursuing hybrid or "lift-and-shift" migrations, fragmented and siloed data continues to impede AI progress ([5]) ([6]). “Enterprises have moved quickly to adopt AI, but adoption alone is not transformation,” said Israel Forst, CEO of Valiantys, emphasising that lasting success requires “trusted data, modern systems, clear governance and an operating model that can turn strategy into execution” ([7]). The survey’s findings reinforce that the race to implement AI must be matched by equal investments in data management and governance fundamentals.
Another new survey underscores how unprepared enterprise data is for AI. Data management firm Komprise published results on 15 October from a September poll of 500 IT and AI decision-makers at large organisations ([1]). While 68% of respondents claimed to have "mature" or maturing AI programmes, only 28% said they have been able to quantify a return on their AI investments so far ([2]). The report exposes a disconnect between perceived progress and the state of underlying data: a staggering 93, 96% of organisations reported that most of their data is not ready for AI and advanced analytics ([3]). The top pain points in AI data readiness were finding the right data for each project and cleaning up poor-quality or duplicate data, followed by challenges in extracting and enriching metadata ([4]).
The Komprise survey reveals that very few companies have done the heavy lifting on unstructured data that AI requires. Only 7% of organisations have cleaned, prepared and classified at least 70% of their data for AI ([5]), and just 4% have integrated more than 70% of their unstructured files into analytics platforms (with appropriate schema), while only 24% have even halfway done so ([6]). These shortcomings in data preparation leave AI projects starved of reliable, accessible information, which in turn limits the business impact of AI. The findings suggest that many firms are still in the early stages of grappling with sprawling "data quality debt," even as they invest in cutting-edge AI tools. Komprise concludes that enterprises must urgently address data cleanliness, integration and governance to turn pilot AI initiatives into scalable, ROI-generating solutions.
An IBM study of 3,200 senior executives across Europe and the Middle East found that 80% had experienced an AI-related incident, ranging from rogue AI behaviour to data exposure, that reached the C-suite or board in the past year ([1]). Despite these high-profile AI failures and compliance scares, fewer than half of affected organisations (just 40%) strengthened their AI governance policies in response ([2]). The IBM report indicates that many companies choose to push ahead with AI projects even after discovering gaps in oversight: 56% of executives said they continued deployments within agreed risk limits while governance “catches up,” and only a third delayed AI implementations until proper controls were in place ([3]). The UK results were starker, 61% of firms kept deploying AI rather than pausing for governance, compared to 33% across EMEA that chose to wait until they were ready ([4]).
IBM’s researchers caution that AI adoption is outpacing the frameworks needed to manage it responsibly. “AI is moving faster than governance, and that gap is becoming a business risk,” warned Leon Butler, Chief Executive of IBM UK and Ireland ([5]). As AI systems take on more autonomous decision-making by 2027, only 37% of EMEA organisations in the study felt their governance structures can scale accordingly ([6]). The study’s findings suggest that without better data governance and risk management, organisations may face growing operational and ethical incidents as they pursue AI at scale.
Even as companies struggle internally with data quality and governance, the external landscape for data in AI is shifting. On 15 October, **Google announced that CNN, the Financial Times, Axios and more than 400 other news outlets have joined its AI-driven news initiative, agreeing to license their content for use in Google’s generative AI features ([1]) ([2]). Under the partnership, an expansion of a programme piloted with a handful of publishers last year, Google will pay these media organisations for the rights to use their journalism as training data and in new AI-powered news products ([3]). Financial details were not disclosed, but the deal marks CNN’s first major agreement since its acquisition by Skydance, as traditional publishers seek compensation for the use of their reporting in AI systems.
This represents a milestone in the evolving economics of data ownership. After years of tech firms scraping news content to develop AI models, prompting legal challenges from publishers, one of the largest AI players is now directly attaching a price to high-quality data and negotiating access on content owners’ terms ([4])**. For enterprise leaders, the move signals a future where having the rights to use trusted external data becomes as important as technology itself. Companies aiming to deploy AI at scale will need clear data partnerships, licensing strategies and compliance checks in place to ensure they have the rights and quality of data required to train reliable AI models.