Agentic Commerce: Shopping and Payments Re-(AI)magined

Source: Inderpreet Batra, Rohan Panjwani, Neema Aggarwal, Ankit Mathur and Alexandria Donohue; Boston Consulting Group · 2025-09

Agentic commerce is poised to disrupt the consumer shopping journey as we know it. Players across the payments value chain will be impacted, and there will be both winners and losers. The time to act is now, while new shopping experiences are still being defined. Those that wait for innovation to stabilize may miss their chance to find a role

Introduction. Over just a few years, the retail sector has seen a fundamental transition, with first digital and then mobile playing an increasingly critical role. In 2024, e-commerce accounted for $2.3 trillion of consumer spending in the US alone. Now, the growth of AI is set to turbo-charge disruption in the industry once more, with agentic commerce offering the possibility of augmentation or even full automation at every step of the shopping journey. Not only that, but consumers are dialed in, with a new BCG survey showing that more than four in five (81%) expect to leverage agentic commerce tools for shopping. On that basis, along with the types of product categories customers find most suitable for agentic commerce, over $1.3 trillion will be impacted by agentic commerce, of which almost $400B will be mostly autonomous. As agentic commerce continues its rise, winners and losers will emerge across the value chain, with consumers and networks standing to gain most, merchants facing fundamental challenges, and issuers presented with both risks and opportunities.

To fully understand what agentic commerce adoption may look like, BCG conducted a survey of 2500 consumers broadly representative of the US population. These are some of our key findings:

• 53% of consumers have already used GenAI tools/chatbots.

• 81% of consumers expect to leverage GenAI tools in shopping for at least one product category.

• 42% of consumers say they would allow a GenAI agent to shop entirely on their behalf for at least one product category.

• If using an AI agent to shop for them, consumers would prefer AI providers that offer loyalty points / cash back, fraud protection, and free delivery as top incentives.

• Consumers cite high levels of familiarity, the ability to return/cancel for a full refund, and lower product cost as the most critical factors when deciding if an agent should have purchase autonomy (See Exhibit 1) – indicating a need to build trust in the technology.

• Based on consumer sentiment – certain product categories emerge as most likely candidates for early agentic commerce adoption: household supplies, ordering food delivery, hygiene and beauty products, over the counter medicine, and various daily use products such as pet food • Consumers believe that retailers, technology firms, and AI-specific companies will make the best agent providers. But they are generally undecided on whether they would adopt one agent or multiple (e.g., depending on product category) across their shopping activities.

Discussion. Our research suggests constraints on the speed and level of consumer uptake of an end-to-end autonomous agent will be driven by both supply and demand factors. On the supply side of the equation, categories or products that are complex, risky, delayed delivery, low online availability, or luxury (e.g., cars, designer jewelry) are naturally likely to lag. From a demand perspective, headwinds may include a lack of customer education around AI agents and low levels of trust initially, exacerbated by news flow in relation to hallucinations and data security. Provider monetization may be a turn off, for example where there are additional subscriptions, fees, or premiums to pay. Therefore, we believe agentic commerce will need to progress through evolutions to overcome these supply and demand considerations.

The e-commerce shopping journey typically consists of four major steps:

1. Trigger: The consumer decides they have a need for a product or service.

2. Discovery: They undertake research to discover the options that meet their needs.

3. Comparison: They compare potential options across product specs, reviews, price, and other factors.

4. Purchase: They make their selection and check out on the merchant website.

Agentic commerce will disrupt each step as it progresses toward full autonomy. We expect the evolutions to overlap as adoption varies by product category, ticket size, merchant, and other factors (See Exhibit 2).

Evolution 1: Consumers trigger the process and embark on discovery and comparison within a large language model (LLM) or AI chat interface such as ChatGPT, rather than the traditional path of using a search engine like Google. From there, they click on a link, often jumping straight to the product or checkout page on the merchant website to complete the purchase.

Evolution 2: As offered by AI-powered answer engine Perplexity Pro, consumers can now select their desired product and check out within the AI interface with a wallet or a stored payment method, never visiting the merchant’s website.

Evolution 3: Agents conduct each stage from discovery and comparison through purchase with minimal to no human input. The agent makes the product selection based on historic preferences and preset criteria and then checks out for the customer with a stored payment method.

Evolution 4: Agents own the entire journey and play a significant role in the purchase step by intelligently optimizing the consumer’s payment methods based on offers and rewards to maximize consumer benefit.

Across these evolutions, we estimate that 55% of e-commerce spend could be agent-assisted and 18% could be driven by autonomous agents; equating to $1.3 trillion (See Exhibit 3).

Consumers will be the biggest winners, with agentic commerce simplifying e-commerce by providing a frictionless shopping experience. They will be able to leverage a personal shopping assistant who understands their preferences and can research, compare, and check out on their behalf across several product categories.

Merchants, on the other hand, are expected to be the most disadvantaged. As agents gain autonomy, their customer relationships and brand relevance will erode, their economics will be disrupted as price transparency rises and they will be subject to higher costs/fees. Often, they will require new marketing strategies to remain relevant. Despite this, there will be opportunities: Merchants may be able to partner with issuers and AI platforms to deliver more personalized offers, and those with niche products may see increased discoverability as search becomes hyper-specific.

Conclusion. Agentic commerce is not simply the next stage of e-commerce – it is a transformation that will redefine the very fabric of how consumers discover, select, shop and pay. The promise of agent-driven ecosystems lies in their ability to deliver to consumers unprecedented levels of personalization, simplicity, and empowerment, while simultaneously unlocking new opportunities for players that are bold enough to adapt. The time to act is now. As agents accelerate in capability and adoption, the players that experiment, partner, and shape the standards of this new ecosystem will capture the advantage – while others risk seeing their relevance fade.