UK firms struggle with shadow AI & weak data controls

Dataiku has published a report that points to widespread unauthorised use of generative AI tools in British businesses, alongside weak data integration and governance gaps that data leaders say undermine AI deployments.

The findings draw on a survey of more than 100 senior British data leaders, alongside anonymous interviews recorded at Dataiku's London Summit. Dataiku reported that respondents raised concerns about employee use of consumer AI tools, the state of underlying data, and the level of technical understanding in executive teams.

Half of senior British data leaders said they believe more than half of employees at their company use generative AI tools for work without permission. The report links this pattern to risks around sensitive information and internal controls.

One anonymous interviewee described frequent use of public tools with company data. "I know almost all of my employees are using shadow AI - and most of them copy and paste sensitive data into ChatGPT every week. I'm guilty of it too," said a senior executive.

Shadow AI

Dataiku framed the unauthorised adoption of tools as a governance issue. The report suggests that staff can move faster than corporate policies and technical controls, particularly where employees use AI for routine work tasks.

The report also found that most data leaders do not see AI as a direct substitute for expert staff. Dataiku said 96% of respondents believe AI systems elevate and enhance expert knowledge rather than replacing it.

An anonymous interviewee described AI use in day-to-day technical work. "In my current role, I need to use SQL every single day, but I'm not even close to being fluent in it. Instead, I rely on AI to help me use SQL and other programming languages, and I know that a lot of my colleagues are the same," said a Data Manager.

Data leaders described this type of usage as widespread, particularly in analytics and engineering functions. The report presented it as one reason why organisations struggle to align policy with how employees actually work.

Data foundations

Alongside workforce behaviour, the report focused on the readiness of corporate data estates. Dataiku said 90% of British data leaders cited data integration challenges or proprietary data access as their top concern when deploying AI. It described this as the highest level reported globally in the wider study.

Several anonymous interviewees said their organisations had started to build AI systems despite unresolved data quality and integration issues. "We are currently building AI randomly, with no regard to the data foundations," said a Senior Architect Engineer.

Another interviewee linked data readiness to business outcomes. "People have invested huge amounts in AI, but the data foundations are so shaky the return on investment won't be there," said a Senior Sales Director.

Data integration issues can include fragmented data platforms, inconsistent definitions for core metrics, and difficulty accessing proprietary sources. The report positioned these factors as practical barriers that slow deployment and increase risk once systems move into production.

Executive gap

The report also flagged a disconnect between technical teams and executive expectations. Dataiku said 78% of British data leaders believe their C-suite underestimates the time and difficulty required to achieve AI that is reliable enough for production deployment.

The report argued that this gap can lead to rushed deployments. It said those deployments can run on "fractured infrastructure" and can increase operational complexity.

Dataiku's Chief Executive and Co-Founder, Florian Douetteau, linked the survey and interviews to a broader governance challenge.

"These confessions are a reminder that AI in the enterprise isn't failing because it's too powerful, but because it's too unmanaged. Behind closed doors, British data leaders are admitting to cutting corners, improvising architectures, and hoping no one notices. That should shake every executive awake," said Florian Douetteau, Chief Executive and Co-Founder, Dataiku.

Douetteau also described what he sees as a competitive dividing line for companies rolling out AI. "The winners won't be the ones deploying the fastest, but the ones building AI on foundations that are governed, explainable, and accountable," said Douetteau.

Dataiku said the report reflects feedback from data leaders at large British businesses. It also said the findings highlight the challenge of applying consistent governance across generative AI, analytics and other forms of machine learning as usage spreads across departments.

Companies in the UK continue to expand AI spending and pilot projects across customer service, software development, marketing and back-office functions. Dataiku said the report indicates that many organisations now face a parallel task of tightening controls, improving data integration, and setting clearer expectations for production-grade deployments.