AI meets the register: ChatGPT is becoming a checkout assistant

Sinziana Albu, Senior News Editor at The Paypers, shares insights into OpenAI’s recent development plans for ChatGPT, with a focus on the service becoming a checkout assistant, the response of the industry regarding its entry in the ecommerce world, and what that means for fraud management, customer protection, and competition in the landscape. 

The development of a payment checkout system with ChatGPT

In July 2025, OpenAI was reportedly planning to integrate a payment checkout system into its chatbot. This initiative aims to enable OpenAI to benefit from an additional source of revenue, capitalising on the heavy traffic of the service besides its traditional subscription model.

At the time of writing, the checkout system is, seemingly, still in development, although OpenAI and its partners have been presenting some early versions of the feature to multiple brands and are in discussion regarding financial terms. When asked to comment on the matter, both OpenAI and Shopify did not respond to requests.

Potential opportunities

The announcement has gathered mixed reactions among several institutions, with some worrying that the system will 'break payments'. For example, the founder and CEO of Kulipa mentioned in a LinkedIn post that the use of agents will affect negatively the payment process in three ways: it will not allow individuals to verify intent anymore, chargebacks will get complicated, and the scaling process will hit differently: 'when agents buying at machine speed with bulk negation power, existing payment infrastructure isn’t built for it'. However, there is potential: if the networks solve the trust problem for agentic payments, they can become essential to the future of commerce.

Officials of Viva.com echoed this forward-looking perspective, mentioning that: 'Conversational commerce is still at an early stage, but it points clearly to how AI will reshape the way transactions happen across digital channels.'

Included in other potential opportunities that are expected to come with this initiative are improved efficiency and cost reduction, as some institutions see the integration of ChatGPT or other AI solutions into the checkout process as an optimised manner to streamline operations and handle tasks like answering customer queries. In addition, the incorporation could also provide clients with an improved experience through continuous availability and personalised services.

Moshe Winegarten, CRO at Ecommpay, mentioned that 'while the noise around agentic commerce, including Chat GPT’s checkout integration, gets louder, we need to pause and examine what's actually changing and what isn't'. He also added that 'the reality is that payment infrastructure is already equipped for agentic commerce. Tokenization exists, authentication frameworks are in place, and PSD3 addresses delegated authentication concerns. From a fraud management perspective, the underlying security mechanisms remain unchanged whether you're clicking buy now on a website or through an AI interface.'

However, potential concerns include the overall loss of human touch, as this process might affect customer satisfaction, since several individuals enjoy interacting with human cashiers for more personalised help.

Privacy, security, and customer protection concerns

The idea of ChatGPT as a checkout assistant has generated concerns about data privacy and compliance. Many in the industry are questioning whether an AI handling sensitive payment and personal information is safe. Institutions and clients alike worry about the risk of data breaches and unauthorised access to financial information.

According to a comment made by the Director of Identity Value Chain Expansion at Mastercard, the traditional identity and fraud stack is currently collapsing, as AI agents begin opening accounts, logging individuals into their accounts, and making payments. She mentions that 'the existing financial payments infrastructure is wholly unprepared for when agents can carry out actions and payments on behalf of humans'.

In a webinar entitled 'From GenAI to Agentic AI: The Future of Travel Payments & Fraud', Aurelie Guerrieri, Chief Marketing and Alliances Officer at DataDome, talks about Agentic AI acting on behalf of customers: 'An AI Agent, by the very nature of how it’s built, looks very much like fraud. This is a bot. It does not behave like a human being. It’s not going to browse the website. When it goes and buys a plane ticket for you, it’s not going to buy you the seat upgrade, it’s not going to check the hotel, it’s going to go straight to the ticket, come out, and maybe it’s going to pay your pricing page 50 times in two days to try and see if it can get a different rate. So, all these rules you have put in place to try and assess what’s human and what’s not are going to get triggered.'

There is also an ambiguity in liability present in this strategy, as it is unclear whether the AI provider or the client will bear the responsibility if an agent misplaces an order or inputs incorrect details. Alongside this, the service might have poor compatibility, as it won’t be able to easily handle additional checkout steps, might struggle with card declines (especially when it comes to international payments), or it may not support alternative payment types like bank transfers or digital wallets.

Chris Hewish, President of Xsolla, mentions that 'ChatGPT stepping into the checkout flow is a meaningful signal that AI is becoming a frontline player in ecommerce. It’s a powerful move, but like all power, it needs guardrails. While conversational AI can streamline discovery and drive conversion, payments aren’t just about speed; they’re about trust. The moment money changes hands is where confidence matters most.'

In addition, the initiative will require retailers to make sure that their AI checkout systems comply with several privacy laws, including GDPR and PCI DSS. Furthermore, the process of implementing AI at the checkout level might require significant infrastructure investments for the retail industry, as companies will face high costs and complexities that are involved in incorporating such systems into their already existing payment processing tools.

However, Moshe Winegarten, CRO at Ecommpay, mentions that ‘the current mixed industry reaction reflects uncertainty about which model will dominate, not concern about payment security. And as agentic commerce evolves, merchants are going to need to ensure their existing payment infrastructure is optimised and observe how these competing approaches evolve in the marketplace.'

According to an article published by Forbes, in order for agentic payments to work, the landscape needs more than just smarter AI; it requires the financial infrastructure to support fast, secure, and non-human transactions, a process that is already falling into place.

This initiative also raised concerns on the topic of competition in the financial landscape. According to an article published by ModernRetail, ChatGPT becoming a pay-to-play marketing platform will make small brands worry that big spenders could dominate product visibility if paid placements are going to dictate recommendations, undermining organic discovery and trust. In the article, Liz Williams, founder of the coat brand The Checkroom, commented: 'I don’t see it being a space that I could compete in, especially against Amazon and H&M and bigger brands. I can see where this would be helpful to companies that already are ranking high on Google, but I don’t see it being something I could afford'.

In addition, several institutions are currently reflecting on which agent model will dominate the industry, and which model will be preferred by customers. Speaking on this matter, Moshe Winegarten, CRO at Ecommpay, said ‘What's fascinating is the emerging battle between different agent models. Some are API-led, which means consumers are locked into a closed ecosystem. Others are browser-based, much like Google is right now. The question is, who will dominate as Agentic Commerce evolves, and what does that mean for the payment choices offered by merchants? And how will consumers react?’ As an example, he explained how Shopify’s API approach pulls merchants into its ecosystem via ShopPay, whereas browser-based agents like Perplexity could work with any stored payment method.

Conclusions

As a long-term outlook, AI-powered checkout assistants such as ChatGPT will likely become more prevalent in the future of ecommerce, especially in digital-first environments, as AI continues to evolve and be leveraged across multiple industries. However, the process of their adoption in physical stores is expected to be balanced with human oversight and personalised customer service to keep customers protected and ensure that the payments are processed correctly and efficiently.

Chris Hewish, President of Xsolla, explained this hybrid approach: ‘AI can support with fraud detection, explain payment flows, and reduce friction. But in nuanced cases such as disputes, digital entitlements, or cross-border compliance, human judgment remains essential. Especially in gaming, where transactions are often irreversible and community trust is everything, we see the future as hybrid: AI as the first layer, humans as the last line of assurance. Bad actors evolve fast. The more capable AI becomes, the more tempting it is to exploit. Prompt injection, misinformation, and synthetic users –  these are real risks. That’s why any checkout experience, especially in gaming, must be designed with both security and empathy at its core.’

At the moment, it's still uncertain whether ChatGPT will become a reliable checkout assistant, especially with the many concerns and questions that are circulating across the industry, but the trend is there: commerce remains a big part of the conversation, as the overall infrastructure of digital payments might face radical changes faster than we expect.

About Sînziana Albu

Sînziana is a Senior News Editor with a keen focus on fintech, payments, and digital identity. With a passion for unravelling the complexities of the rapidly evolving technological landscape, Sînziana is dedicated to delivering insightful news that keeps her readers informed.