AI agents tipped to reshape enterprise eCommerce by 2027

Logicbroker has published a survey of more than 600 enterprise eCommerce leaders suggesting that artificial intelligence agents could influence a growing share of online purchases within the next year and account for a large proportion of transactions by 2027.

The report, titled The State of Agentic Commerce Adoption, found that more than 90% of respondents expect AI agents to influence at least 20% of online orders by 2027. More than one-third said AI could shape more than half of all online transactions by then.

Expectations for payback also appear to be shortening. Nearly half of respondents anticipate a significant return on AI investments within the first year-faster than typical technology investment horizons in large organisations.

Logicbroker describes its work as "agentic commerce": software agents that go beyond recommending products to taking actions that affect the transaction itself. This can include selecting products, choosing suppliers, and placing orders through integrated systems.

"The conversation around AI in commerce has focused heavily on discovery-think chatbots, search assistants and product recommendations," said Omar Qari, CEO of Logicbroker. "But what the data is actually suggesting is that AI is starting to move deeper into the transaction itself. When software agents begin deciding what gets purchased and where it gets sourced from, the structure of digital commerce starts to change."

Spend and timelines

The survey points to broad deployment: 95% of enterprises surveyed have already rolled out at least one AI-driven commerce feature. Almost half also plan significant spending over the coming year.

In total, 47% of organisations intend to invest USD $1 million or more in AI-driven commerce initiatives in the next 12 months; within that group, 21% expect to spend more than USD $5 million.

Respondents expect those investments to deliver quickly. Three in four anticipate return on investment within 24 months, and close to half foresee returns within the first year.

Deployment schedules also appear to be accelerating. More than half of organisations plan to roll out AI shopping agents in the next six months.

Beyond retail

The findings suggest the shift is not limited to consumer retail. Hybrid organisations with both B2B and B2C operations were the largest category of respondents at 45%, while pure retail companies accounted for 22%.

This split matters because B2B buying processes often involve negotiated pricing, contract terms, supplier compliance checks, and structured approval workflows. Those factors add complexity when companies try to automate parts of purchasing with AI agents.

Ed Bradley, Chief Growth Officer at Virtualstock, powered by Logicbroker, linked agent activity to changes in customer acquisition models and product discovery.

"Legacy retail organizations that aren't prepared for AI will feel the disruption first," Bradley said. "As AI agents increasingly discover, compare and purchase products autonomously, traditional SEO and paid acquisition channels will face structural pressure and will need new technologies to overcome those challenges."

Integration focus

Many enterprises appear to see implementation and integration as the main work ahead, rather than building their own foundation models. Fewer than 15% of respondents said they are developing proprietary large language models.

Many are also adopting several commercial AI platforms at once, increasing the need to integrate AI tools with core commerce infrastructure such as product information, order management, inventory, pricing, fulfilment, and supplier systems.

Integration was cited directly as a constraint: four in 10 respondents said better integration tools would speed up AI adoption.

Executive hesitation was not a major obstacle in the sample, with only 12% citing leadership buy-in as a barrier. More pressing challenges included systems connectivity, data quality, and integration complexity when moving from pilots to scaled deployments.

Bradley described a maturity path that moves beyond decision support. "Longer-term initiatives, including AI-powered supplier management and autonomous reordering, signal the next stage of maturity, where AI will move beyond supporting human decisions to executing transactions independently," he said.

UK relevance

Although the survey focused on US enterprises, it offers a read-across for UK retailers and brands operating similar technology stacks and supply chain models. Many large UK organisations span marketplaces, direct-to-consumer sites, and B2B channels, supported by third-party fulfilment and multi-supplier catalogues.

The report also suggests most organisations are not trying to be early pioneers. Only 18% of respondents described themselves as industry leaders breaking new ground, while the rest appear to be moving once approaches are proven and implementation risk is manageable.

Qari framed the competitive question around connecting platforms rather than building them.

"The enterprise AI race isn't about who builds the best model. It's about whether companies can connect multiple AI platforms to the infrastructure that actually powers commerce. In a multi-model market, betting on a single provider is a risk," Qari said.