Retailers are now entering a more cautious phase in their relationship with AI‑driven commerce.
After a period of rapid experimentation and high expectations, many brands are now reassessing how deeply artificial intelligence should be embedded into the buying journey, particularly when it comes to conversion and checkout.
While AI has proven effective at reducing friction and accelerating discovery, its growing role as an intermediary between brands and customers has raised concerns around data ownership, trust, and long‑term customer value.
As a result, retailers are beginning to scale back certain integrations, seeking a more balanced approach that preserves the speed and relevance AI offers without surrendering control over the customer relationship.
ChatGPT Checkout Exposes Gaps in AI‑Driven Commerce
In September 2025, Shopify announced its decision to partner with OpenAI to enable AI‑driven product discovery and purchasing experiences within ChatGPT, positioning conversational agents as a new entry point to online storefronts.
By letting consumers discover, decide, and purchase within a single conversational flow, ChatGPT became a point of sale.
For retailers, this new approach promised faster conversions and reduced drop‑off rates, and enabled OpenAI to enter the retail world with a new revenue stream.
However, just months after their announcement, OpenAI confirmed its decision to retreat from frictionless AI checkout, having underestimated the complexity of real‑world commerce and the trust signals consumers expect at the point of purchase.
As early-onset enthusiasm gave way to caution, retailers are now re‑evaluating the risks, rewards, and long‑term implications of embedding third‑party AI directly into the buying experience.
When Convenience Comes at the Cost of Control
As a result, the trade‑off between convenience, automation, and ownership of the customer journey has now moved from a theoretical concern to an operational reality for retailers.
With the promise of speed and simplicity after enabling in‑chat product discovery and purchasing inside ChatGPT less realistic than expected, retailers were forced to reassess how much of their commerce stack they were willing to hand over in exchange for convenience.
When the buying moment shifts into a third‑party AI environment, retailers begin to lose control over the most valuable parts of the customer relationship, and what initially appeared to be a distribution win instead exposed deeper concerns around ownership, trust, and long‑term value.
Speaking with CX Today, Matt Wurst, CMO at Genuin, argues that in-chat checkout ultimately strips away control over the brand experience and customer relationship, reducing retailers to interchangeable products within a platform they do not own.
“The honeymoon’s over. Retailers signed up for distribution and are realizing they traded the customer relationship for it,” he explained.
“In‑chat checkout abstracts away everything that makes a brand a brand: merchandising, the post‑purchase moment, the data, the feedback loop. You’re an SKU in someone else’s interface.”
When discovery and checkout happen outside a brand’s owned channels, retailers can lose direct visibility into customer behavior, therefore limiting access to first‑party data, weakening personalization and loyalty strategies, and likely losing opportunities to shape post‑purchase relationships.
The checkout moment is where brands capture the richest insights, reinforce trust, and outsourcing that moment risks turning a customer relationship into a one‑time transaction mediated by someone else’s platform.
As a result, retailers are increasingly drawing clearer boundaries around where automation should accelerate the journey and where ownership of the experience remains non‑negotiable.
Why Outsourcing Checkout Creates Long‑Term Exposure
With these concerns now at the front of many brands’ minds, retailers are hyper-focused on the concrete risks of relying on third‑party AI platforms for conversion and transaction capture.
When customer interactions are mediated through external AI assistants or marketplaces, the platform often becomes the primary interface between the brand and the buyer, ultimately shifting ownership of the conversational journey and limiting a retailer’s ability to understand how intent is formed.
As a result, the brand risks becoming a supplier inside someone else’s experience rather than the owner of a direct customer relationship.
Data and intelligence risks also enhance this problem, as customer conversations with high value signals take place inside third‑party AI environments, retailers may receive only partial visibility or summarized outputs, rather than full access to raw interaction data.
This can cause the external platform to accumulate a deeper understanding of customer behavior than even the retailer itself, weakening the brand’s ability to build intelligence or improve lifetime value.
Elissa Brown, E‑Commerce Lead for North America at AppsFlyer, told CX Today that handing checkout to third‑party AI platforms strips retailers of visibility into conversion and customer behavior.
“The checkout moment is the most data‑rich, relationship‑defining interaction. If you can’t see conversion, you can’t learn,” she warns.
“Without learning, you can’t build loyalty – and without loyalty, you’re renting customers from an LLM.”
This can also increase operational and financial dependency, where changes made to APIs, fees, or compliance requirements by third-party AI platforms can affect conversion performance with little warning.
For retailers, as a result, switching providers can be complex, costly, and disruptive, reducing strategic flexibility and increasing exposure to decisions made outside their control.
Furthermore, this can be complicated by measurement and attribution challenges, where retailers may lose clarity on which interactions drove conversion if discovery, conversation, and checkout occur within third‑party AI interfaces.
This can make it harder to optimize funnels, allocate spend effectively, or understand true channel performance, whilst governance and compliance risks grow as sensitive customer data flows through systems the retailer does not directly operate or govern.
The Engagement Risk of Fragmented AI Strategies
As retailers pull back from third‑party AI checkout and transaction capture, retailers must now combat the challenge of ensuring that reduced integration at the point of sale does not translate into fragmented engagement elsewhere in the customer journey.
In conversation with CX Today, Rodney Mason, CMO at Minty, argues that avoiding AI entirely risks losing visibility and relevance at the discovery stage.
“Scaling back AI at the checkout stage makes sense, but scaling back entirely does not,” he explained.
“Brands that aren’t showing up in those conversations are invisible at the moment a decision is being made.”
Without shared context, AI‑driven responses may feel generic, forcing customers to repeat information or navigate inconsistent experience channels, directly undermining engagement and increasing effort.
Personalization strategies are also vulnerable, as effective personalization depends on AI’s ability to access real‑time context, behavioral signals, and historical interactions across the full journey.
When AI operates in silos, personalization becomes shallow and reactive, meaning offers, recommendations, and support may feel misaligned, mistimed, or exhibit a lack of awareness, meaning brands miss opportunities to create value earlier in the decision- making process.
These gaps can erode loyalty, and if AI cannot maintain conversational memory, support seamless handoffs, or reinforce post‑purchase relationships, the experience becomes less relational.
Though customers may still complete purchases, trust and long‑term affinity weaken, making switching easier in a competitive market where expectations for personalized, always‑on engagement continue to rise.
Agentic Commerce Enters Its Reality Check
As retailers refine how and where they deploy AI, the conversation is shifting toward the future of agentic commerce, a model in which AI agents do more than answer questions or surface recommendations.
In agentic systems, AI actively participates in the buying journey by understanding intent, comparing options, completing tasks, and occasionally initiating actions on a customer’s behalf.
Eventually, AI is expected to play a central role in how consumers discover products, evaluate choices, manage subscriptions, reorder essentials, and receive post‑purchase support.
This also has the potential to make commerce faster, more contextual, and more personalized, reducing the effort required to move from intent to outcome.
As Tiffany Johnson, CPO at NMI, explained to CX Today, agentic commerce still holds significant potential, but only if AI earns consumer confidence and gives merchants assurance over how their products and demand are represented.
“OpenAI has stepped back from instant checkout, but this doesn’t mean agentic commerce is stalling. It feels more like a reset,” she said.
“Consumers need confidence the AI is acting in their best interests, and merchants need assurance that their products are represented accurately and that demand is genuine.”
Agentic commerce creates the most value in moments that benefit from speed, relevance, and decision support, relevant to several areas where AI can meaningfully reduce friction.
By remembering preferences, understanding past behavior, and responding in real time, AI can help customers narrow choices and make more confident decisions, improving engagement earlier in the journey while lowering service costs and increasing personalization.
Despite this, the future of agentic commerce depends on clear boundaries, meaning AI should not fully own moments that require emotional trust, significant financial commitment, or complex personal judgment.
For high‑value purchases, sensitive financial decisions, regulated services, and emotionally driven buying moments that demand human reassurance, AI works best as an enabler rather than as the sole decision‑maker.
How Retailers Are Reclaiming Control Without Losing AI’s Benefits
As retailers recalibrate their AI strategies, the best practices are now focusing on selective deployment rather than full automation.
When AI is able to improve speed, relevance, and operational efficiency without removing the brand from critical customer interactions, retailers are more likely to see strong returns from using AI in areas where automation reduces friction and enhances personalization while keeping the brand visible and accountable.
At the same time, higher‑value moments in the customer journey require stronger brand ownership.
Competitive retailers are keeping these moments within their own platforms, ensuring transaction and relationship data remain first‑party assets, while also prioritizing AI experiences that reflect brand voice, merchandising strategy, and service standards.
Consistency across channels is key, as customers increasingly expect seamless recognition regardless of where they engage.
Choice is also emerging as a best practice, where by giving customers clear options between automated and human support helps maintain confidence.
Rather than replacing human interaction, AI is being used to prepare context, reduce wait times, and support staff with better insights, resulting in a balanced model where AI enhances convenience and intelligence while retailers retain control over the moments that define brand value, loyalty, and lifetime revenue.
“The brands that win are the ones that show up in AI‑driven discovery without giving up the transaction,” Mason added.
“Map the AI‑assisted journey end‑to‑end, then decide intentionally where AI belongs, and where it doesn’t.”
