AI in Retail: The cost vs cost benefit conundrum

AI adoption continues to grow across UK retail

Amid growing noise around AI, it can be difficult to distinguish the headline‑grabbing narratives from the more grounded realities.

Adoption across the UK retail sector has been steady, with 50% of retailers already using AI (Knight Frank, 2023) but its path is far from uniform.  The technology has clear potential to reshape how retailers operate.

However, the sector continues to weigh this potential against tight cash flow, high capital requirements and a broad lack of confidence in tangible returns. The result is a trajectory shaped as much by hesitation as by interest.

Why retailers are investing in AI

UK retail already operates under significant costs pressure and intense competition. These conditions have created a strong incentive to explore AI.

In 2025, UK businesses increased investment into AI-enabled technologies by 73% year-on-year, and 48% cited operating cost pressure as a key adoption driver (RTIH, 2026 & Eversheds Sutherland, 2025). 

However, enthusiasm alone does not directly translate into deployment. Retailers remain constrained by the substantial initial costs of AI investment and the pressure these exert on already stretched operating budgets. Although 67% report having ‘clear AI investment plans’, 40% state they lack the financial capacity to progress strategies. Many are waiting for evidence of significant measurable gains and remain unconvinced by the current ROI profile. This scepticism is reinforced by mixed real-life outcomes.

Where AI is delivering the greatest value for retail

High‑profile experiments have captured attention but have not always delivered results.

Ocado’s pausing of its automated warehouse rollout internationally and Amazon Fresh / Go store closures highlighted the gap between theory and practice. This dynamic may create a two-tier adoption curve in which larger operators lead the innovation charge, whilst smaller retailers only pick up the tools that prove quantifiable operational advantages.

The most transformative effects are expected in areas least visible to consumers, such as stock management, forecasting accuracy and labour allocation. Data improvements will help retailers juggle inventory in real time and curate product ranges more effectively.

This increased insight also helps smooth the flow of goods across both physical and online channels. As stock accuracy improves, the need for large back of house areas may decline, creating modest potential for freed-up trading space and gradual recalibration of operational space. 

AI at the retail customer interface

This operational backdrop also shapes how AI is being applied at the customer interface.

Zara’s recent virtual fitting functions, which increased click through rates by 18% and cut shoot costs by 35%, indicate that AI’s strongest value could be delivered through incremental gains (City AM, 2026). While human validation remains essential, the workforce composition will likely evolve.

By 2035, two thirds of support and supply tasks and more than 70% of digital and technology operations are forecast to be AI-enabled, bringing further cost benefits. As data access improves, AI will allow retailers to build a clearer picture of customer behaviour, product interest and local demand patterns.

This could translate into more informed occupiers who are able to take faster and more confident decisions around where to locate and invest in assets. Retailers operating more efficiently are generally more resilient.

What AI means for retail real estate

This has positive real estate implications. This occupier strength can help them to remain in leases for longer, generate stronger cash flow. It in turn enhances their capacity to sustain rental levels, putting the occupational market in good health and acting as a catalyst for rental growth.

AI is best at play as a supplement to established retail disciplines rather than a replacement, and reinforcing stronger fundamentals, rather than redefining them. The most important consequence for real estate is therefore likely to be a more durable and better performing occupier base, rather than immediate or dramatic change to the configuration of retail space.