Two of the most popular AI chatbots are heading to the checkout line. OpenAI and Perplexity have released shopping-focused AI features designed to make it easier to discover products, compare prices, and purchase items directly within chat. The rollout comes just as holiday shopping — and impulse spending — surges. While tech giants expand into commerce, emerging competitors remain unfazed, betting that deep vertical expertise and proprietary data will set them apart.
What the New Assistants Offer Shoppers
OpenAI’s successor to ChatGPT focuses on guided product research. Users can describe constraints — for example, a gaming laptop under $1,000 with a 15-inch screen — or even upload an image of an outfit and request cheaper alternatives. The goal is to fuse browsing and shortlisting into one seamless conversation.
Perplexity takes a different angle, using persistent memory to personalize recommendations. The system remembers prior preferences and contextual details — climate, commute routines, job requirements — to narrow choices without repeated prompting. Both companies are also streamlining checkout: OpenAI integrates with partners like Shopify for native purchases, while Perplexity connects to payment systems such as PayPal for in-chat transactions.
This timing is intentional. Adobe forecasts a 520% rise in AI-powered shopping interactions this holiday season, signaling growing consumer trust in bots for purchase advice. With broad reach and brand familiarity, both companies are well-positioned to capitalize on that momentum.
Why Startups Aren’t Afraid of AI Giants
Industry experts argue that general-purpose chatbots still lack merchandising nuance. According to niche players like Phia, Cherry, Onton, and Deft, real performance depends less on raw language models and more on rich, structured retail data — clean catalogs, standardized product attributes, and real-time updates on pricing and availability.
Onton’s team, for instance, built a robust pipeline that categorizes hundreds of thousands of furniture SKUs using unified taxonomies, image-quality checks, and detailed style metadata. Such curation enables fine-grained filters — like arm style or seat depth — that generic assistants often reduce to vague suggestions. As Onton CEO Zach Hudson explains, startups relying purely on off-the-shelf models will struggle, but those with domain-specific data and workflows can thrive.
The strategic bet: vertical assistants can convert more effectively by learning the trade-offs shoppers face in complex categories such as fashion, travel, or home goods. This creates healthier unit economics even without the vast traffic of mainstream platforms.
Data and Checkout: The Real Battleground
Scale still matters. OpenAI and Perplexity can forge partnerships that integrate discovery and payment in a single experience. Native checkout drastically reduces user drop-off compared to redirecting shoppers to external retailer sites. Smaller startups, however, often spend months securing similar deals with platforms like Shopify or payment processors — frequently at higher costs or volume commitments.
Yet specialization brings its own advantages: first-party fit data, return feedback, and consistent attribute frameworks that power superior recommendations on size, compatibility, and product bundles. In detailed categories — from tech specifications to furniture dimensions — these advantages lead to small but meaningful gains in conversion and reduced buyer’s remorse.
The Monetization Blueprint
Neither OpenAI nor Perplexity has a fully realized profit model, and compute is expensive. Expect their e-commerce playbook to mirror search engines: affiliate fees, merchant integrations, and sponsored placements within results. Retail media is already one of the fastest-growing ad channels, and analysts from Insider Intelligence and Forrester anticipate more growth as retailers monetize high-intent searches.
However, the familiar pitfalls remain. If paid placements overshadow organic results or lack clear labeling, users may grow frustrated — as they did with cluttered web ads. Transparent disclosure, relevance control, and fair return policies will determine whether conversational shopping builds trust or repeats old mistakes in a new format.
Testing Ground: Early Results from AI Shopping
Early case studies suggest promise. Klarna’s AI assistant for customer service and product questions has sharply reduced human workload and improved resolution times. These outcomes hint that shoppers will embrace AI support when it’s accurate, fast, and accountable.
Accuracy, however, remains the core challenge. Hallucinated specs, outdated prices, or missing sizes can translate into costly returns and complaints. To prevent that, major players are developing product knowledge graphs, verified retailer data feeds, and live price-monitoring systems. Privacy remains equally vital: assistants that track preferences must give users clear options to consent, delete, or modify stored data, in line with regulatory guidelines.
What to Watch as AI Shopping Goes Mainstream
- Three performance metrics will define success:
- The share of chatbot sessions that result in purchases without leaving the chat.
- Conversion and average order value compared to traditional category benchmarks.
- A reduction in returns linked to poor recommendations.
McKinsey estimates generative AI could unlock $400–$660 billion in value across retail and consumer goods. Those able to turn conversational engagement into profitable, low-friction sales will capture a major share of that opportunity.
For now, OpenAI and Perplexity bring reach and transaction capability, while startups counter with depth and data precision. The game won’t be zero-sum — expect partnerships, data-sharing deals, and white-labeled solutions to emerge. The upcoming holiday season will reveal whether general AI chatbots can truly shop or if, in the end, specialists still hold the upper hand where it matters most: at checkout and beyond.
