39% of retailers expect AI to exceed 10% of tech spend by 2028

Nearly two in five retailers expect artificial intelligence to command more than 10% of their technology budgets by 2028, signalling a significant shift in enterprise spending priorities as the industry moves beyond experimentation towards scaled deployment. The finding, drawn from December 2025 research by the National Retail Federation, underscores retailers' growing confidence in AI's business case even as many grapple with implementation challenges and unproven returns.

How much are retailers currently spending on AI?

Current AI spending remains relatively modest across most retail organisations, though precise figures vary widely by company size and digital maturity. Industry observers note that many fashion retailers are still in pilot phases, allocating small percentages of IT budgets to AI projects whilst they evaluate use cases and build internal capabilities. Early adopters have begun to increase allocations as proof-of-concept initiatives demonstrate tangible value.

The NRF's December 2025 figure of 39% expecting AI to exceed 10% of tech spend by 2028 represents a notable acceleration from current levels, suggesting retailers anticipate moving from selective pilots to broader platform investments over the next two years. This timeline aligns with typical enterprise technology adoption curves, where initial caution gives way to more aggressive spending once risk and reward profiles become clearer.

Which AI use cases are retailers prioritising?

Fashion retailers are concentrating their AI investments in areas with clear operational impact and measurable returns. The most frequently cited priorities include:

• Demand forecasting and inventory optimisation, helping retailers reduce overstock and stockouts whilst improving cash flow
• Personalisation engines that tailor product recommendations, marketing messages and shopping experiences to individual customer preferences
• Supply chain visibility and logistics optimisation, particularly for managing complex global sourcing networks
• Customer service automation through chatbots and virtual assistants that handle routine enquiries
• Visual search and product discovery tools that allow shoppers to find items using images rather than text

Demand forecasting has emerged as an especially attractive starting point because it addresses a persistent retail pain point whilst leveraging existing transactional data. Poor demand planning costs retailers billions annually through excess inventory that must be marked down or written off, creating both financial incentive and clear metrics for AI success.

What ROI are retailers expecting from AI investments?

Return on investment expectations fall into two broad categories: cost reduction through automation and efficiency gains, and revenue growth through improved customer experience and conversion. On the cost side, retailers anticipate AI will reduce labour requirements for routine tasks, lower inventory carrying costs through better forecasting, and decrease markdown rates by aligning assortments more closely with demand.

Revenue expectations centre on personalisation's ability to increase average order values and purchase frequency, as well as improved conversion rates from better product discovery and customer service. However, quantifying these benefits remains challenging, particularly for revenue-focused applications where attribution is complex and results may take time to materialise.

The gap between expectation and reality has proven significant for some early implementers. Retailers report that AI projects often take longer to deploy and require more organisational change than initially anticipated. Data quality issues, integration challenges with legacy systems, and the need to retrain staff can all delay time to value and dampen returns.

What barriers are slowing AI adoption in fashion retail?

Despite growing investment appetite, several obstacles continue to constrain AI deployment across the fashion sector. Data infrastructure remains a primary challenge, as many retailers lack the clean, well-organised datasets that AI systems require to function effectively. Fashion's highly seasonal and trend-driven nature also complicates matters, as algorithms trained on historical data may struggle to predict demand for novel styles or sudden trend shifts.

Skills shortages represent another significant barrier. Retailers compete with technology firms and other industries for data scientists and machine learning engineers, often at a disadvantage given salary expectations and the perception of retail as less cutting-edge than pure technology companies. This talent gap forces many retailers to rely on third-party vendors rather than building capabilities in-house, raising questions about long-term competitive differentiation.

Cost concerns also temper enthusiasm, particularly among mid-sized retailers with limited IT budgets. Whilst AI vendors increasingly offer software-as-a-service models that reduce upfront investment, the total cost of ownership including data preparation, integration, change management and ongoing optimisation can still prove substantial. Retailers must weigh these costs against uncertain returns, a calculation that becomes more difficult when executive teams lack technical expertise.

How will AI spending reshape retail technology budgets?

If 39% of retailers do indeed allocate more than 10% of technology budgets to AI by 2028, the shift will necessitate trade-offs elsewhere in IT portfolios. Legacy system maintenance, traditional analytics tools and some custom development projects may see reduced funding as organisations redirect resources towards AI platforms and the infrastructure required to support them.

This reallocation could accelerate the broader modernisation of retail technology stacks, as AI applications often require cloud infrastructure, real-time data pipelines and API-based architectures that legacy systems cannot easily support. Retailers may find themselves compelled to retire or replace outdated systems sooner than planned simply to enable AI use cases, creating a virtuous cycle of technology renewal.

The competitive dynamics of AI spending also merit attention. If a significant minority of retailers substantially increase AI investment whilst others lag, performance gaps may widen between leaders and laggards. Those who successfully deploy AI for demand forecasting and personalisation could enjoy structural advantages in margins and customer retention, potentially forcing competitors to respond with their own accelerated investments regardless of immediate ROI clarity.