UK firms boost AI privacy but lack strong training & oversight

A majority of UK businesses have strengthened their privacy measures following the adoption of artificial intelligence, but only a minority report making substantial changes, according to new research from technology firm Zoho.

The survey found that 63 per cent of UK organisations have enhanced their privacy protocols after introducing AI. However, just 34 per cent of those organisations describe these enhancements as significant. The study highlights a largely reactive approach, with most businesses strengthening privacy only after experiencing issues related to AI implementation.

Survey scope

The research, commissioned by Zoho through Arion Research, was based on responses from 363 UK business leaders representing small, mid-market, and large enterprises. It explored key themes such as AI adoption rates, privacy practices, workforce readiness, and the maturity of governance frameworks.

Privacy and security remain prominent concerns for many British businesses contemplating AI adoption. The survey reports that 37 per cent of respondents view these issues as barriers to implementation. Customer data breaches were identified as the primary concern, with 44 per cent ranking them as their leading worry.

To address these concerns, 85 per cent of the organisations surveyed now employ dedicated privacy officers or teams. While this indicates a commitment to safeguarding customer data, the report suggests that focusing solely on privacy infrastructure – a 'fortress' mentality – may neglect challenges such as algorithmic transparency, AI bias, and the handling of training data.

Training and skills gap

Developing the necessary workforce skills continues to pose difficulties. According to the survey, only around 20 per cent of businesses have formally trained between 0 and 10 per cent of their workers on AI-related practices. Just 8 per cent have trained between 90 and 100 per cent of their employees, indicating a dependence on informal learning methods for most companies.

Respondents identified their highest upskilling priorities as data analysis (56 per cent), prompt engineering (44 per cent), and building general AI literacy (37 per cent). Additionally, 32 per cent of survey participants cited a lack of technical expertise as a barrier to adopting AI in their operations.

Governance and transparency

Challenges remain in terms of governance structures and transparency. The report notes that fewer than half of organisations have a documented policy for the use of AI (48 per cent) or requirements outlining explainability for AI-driven decisions (45 per cent). These governance gaps, the study notes, can undermine both accountability and trust.

It is still early in terms of business AI adoption but it is expected to become more pervasive. However, organisations need to make step changes in order to mitigate risk and use AI effectively to increase positive impact. This research shows that while many UK organisations are strengthening privacy measures, too often these steps are taken reactively rather than through forward-looking planning.

Sachin Agrawal, Managing Director, Zoho UK, continued, "It's encouraging that most businesses have dedicated privacy officers in place, but without the right training, governance, and clear AI use policies, these efforts may not translate into meaningful protection. To unlock AI's full potential, organisations must go beyond compliance, embedding transparency, investing in workforce skills, and building well-defined data strategies that both safeguard information and maintain customer trust."

AI adoption trends

The Zoho survey points to active adoption of various AI technologies in the UK, spanning traditional AI, generative AI, and agentic (autonomous) AI systems. Generative AI leads, with high usage rates for customer support (67 per cent), content creation (66 per cent), and coding or development assistance (62 per cent).

Traditional AI is also well-established in business settings, with predictive analytics (64 per cent), customer segmentation (63 per cent), and recommendation systems (60 per cent) seeing widespread adoption. Agentic or autonomous AI, though newer, is gaining ground in areas such as personalised customer journeys (58 per cent), autonomous decision-making and robotic process automation (61 per cent), and strategic planning assistance (63 per cent).

Most UK businesses are taking a staged, strategic approach to AI, often starting with individual use of public language models and progressing to broader, organisation-wide integration. Respondents favour a hybrid sourcing model that combines off-the-shelf commercial solutions with custom-built applications, either developed in-house or with partners.

Survey participants indicated their chief areas of AI investment as embedding AI into enterprise systems, developing custom or hybrid solutions, and in-house development. Application priorities include customer service (43 per cent), software development and coding (41 per cent), fraud detection (34 per cent), automation (31 per cent), marketing (30 per cent), and product development (29 per cent).