Artificial intelligence has moved beyond a futuristic promise and become a tangible reality in the retail sector. Recent industry reports, such as Salesforce's Agentic Enterprise Index 2026, reveal that companies are tripling the number of AI agents deployed, with a 53% reduction in deployment time. This advancement is not only transforming how companies interact with their customers but also redefining operational efficiency in a sector as competitive as retail.
Retail has become the epicenter of this technological revolution , accounting for 22% of the total volume of automated work (measured in Agent Work Units, AWU). This figure is no coincidence: retail handles a massive volume of repetitive interactions, from order inquiries to shipment tracking, which lend themselves perfectly to automation. However, the adoption of this technology is not without its challenges, especially regarding systems integration and preserving the direct relationship with the customer.
Retail leads the adoption of AI agents
The data from Salesforce's 2026 Agentic Enterprise Index, compiled from information gathered by organizations that kept their agents in production between February 2025 and April 2026, is compelling: the average number of activated agents per company has nearly tripled in the last year. This rapid adoption is partly explained by the reduction in deployment time, which now averages two days from agent creation to deployment.
The retail sector stands out in this landscape. The volume of AI Users (AWUs) in retail has increased eighteenfold during the analyzed period , and although the sector is characterized by agents that perform one or two specific actions, such as order inquiries or shipment tracking, their adaptability during peak demand is remarkable. A clear example is the holiday season, where an AI agent in retail can operate across nine different skills, a 350% increase compared to its typical configuration.
From conversation to action: the new role of agents
The evolution of AI agents has progressed from simple conversation to the execution of complex tasks. According to a Salesforce report, the ratio of action calls to exit tokens has grown at a compound monthly rate of 15%, indicating that agents are performing more actual work rather than just generating text. This execution capability is especially relevant in retail, where an agent can review an order, apply a business rule, and issue a refund within the same workflow.
A notable example is Pandora, the Danish jewelry company, which has implemented an agent called Gemma to handle routine inquiries and offer recommendations during peak periods. The company attributes the management of 60% of requests and a 10% improvement in its Net Promoter Score (NPS) during the analyzed peak periods to this system. While this is a specific example, it demonstrates how AI can generate tangible value in the retail sector.
Integration, the pending challenge for retail
Despite the progress, the adoption of AI in retail is not without its obstacles. GlobalLogic, a Hitachi Group company, warns about the "Bot Zoo" phenomenon, where multiple automated solutions coexist without communicating effectively with each other. This scenario, according to Gabriel Arango, Head of Technology for Latin America at GlobalLogic, creates a complex technological architecture where employees end up acting as a bridge between disconnected systems.
A common example is returns: a customer starts the process through the mobile app, but upon arriving at the physical store, the point of sale doesn't have access to that information, forcing them to repeat the process. The lack of integration between channels remains one of the main challenges to offering a truly seamless omnichannel experience.
Sales and productivity: the real impact of AI
The data on the impact on sales is revealing. During the 2025 holiday season, retailers that employed AI agents saw an 8% year-over-year increase in online sales, compared to 2% for those that did not. This fourfold difference in growth rate suggests that AI not only improves operational efficiency but also has a direct impact on business results.
Internally, agent adoption is also growing. The weekly frequency of employee interactions with agents increased by 300% , and on platforms like Slack, these systems reached an average of 67 sessions per week. Salesforce estimates that its workforce saves up to five hours per week per employee thanks to the automation of routine tasks and human-free sales processes.
For retailers, the next step will be finding the balance between automation and preserving the direct relationship with the customer. As Adyen points out, the risk of AI controlling purchase intent could weaken what retailers have taken years to build: the direct relationship with their shoppers. The key will be for technology to enhance, not replace, that relationship.
In short, artificial intelligence is redefining the retail sector in Spain and Europe, with significant advances in efficiency and sales. However, long-term success will depend on companies' ability to integrate these technologies coherently, preserving customer trust and maximizing the potential of AI agents.
