AI Shopping Assistants in 2026: How Goddiva, THG Ingenuity, and Walmart Are Transforming Retail — and Why eBay Risks Falling Behind

An AI shopping assistant is a conversational, AI-driven software agent integrated into an e-commerce platform that helps users discover products, track orders, process returns, and receive personalized discounts through natural language interaction—a shift from rigid, rule-based chatbots to large language model (LLM)-powered experiences. In 2026, major retailers are deploying increasingly sophisticated versions, while others risk alienating shoppers with poorly designed implementations.

Goddiva AI: A Fashion-First GPT-5 Assistant

UK-based fashion e-commerce brand Goddiva has launched "Goddiva AI," a shopping assistant powered by OpenAI’s GPT-5 and the LangChain framework retailtechinnovationhub.com. Integrated directly into the brand’s website and backend order systems, Goddiva AI offers real-time order tracking, personalized discount prompts, and natural language interaction for product discovery, order support, and returns. The company describes the experience as aiming for a "personal, intuitive, and fun" outfit discovery journey.

The key change is that Goddiva AI moves beyond simple FAQ bots. Because it is built on GPT-5—one of the most advanced models available—the assistant can hold context over multi-turn conversations, remember a shopper’s style preferences from previous interactions, and even proactively suggest complementary items. For example, a customer browsing a wedding guest dress can ask "Do you have this in emerald green?" and immediately receive availability, price, and a matching accessory suggestion—all without leaving the chat interface. This level of integration signals that fashion e-commerce is beginning to treat AI not as a gimmick but as a core sales channel.

THG Ingenuity and Google Cloud: Scaling AI for Enterprise Retail

THG Ingenuity, the e-commerce technology and services division of THG (The Hut Group), has unveiled a new AI-powered shopping assistant developed in collaboration with Google Cloud retailtechinnovationhub.com. This partnership is significant because THG Ingenuity powers the online stores for hundreds of brands across beauty, nutrition, and luxury goods. By embedding Google Cloud’s AI capabilities—including Vertex AI and Gemini models—the assistant can deliver personalized product recommendations, handle complex natural language queries, and scale to millions of concurrent users.

Feature	Goddiva AI	THG Ingenuity + Google Cloud	Walmart Sparky
Core model	OpenAI GPT-5	Google Gemini (via Vertex AI)	Proprietary (likely fine-tuned LLM)
Primary use case	Fashion outfit discovery and order support	Enterprise retail search and recs	Live shopping deals and interactivity
Key differentiator	Real-time order tracking	Multi-brand scalability	Integration with live video streams
Deployment date	November 2025	June 2026	June 2026
User reaction	Positive (early reports)	N/A (enterprise-facing)	Mixed (some excitement, some confusion)

For retailers using THG’s platform, this assistant means they can offer a level of AI sophistication previously reserved for tech giants like Amazon. A shopper looking for a "vegan protein powder with chocolate flavor, under $40" can type that exact phrase and receive curated results, complete with ingredient comparisons and user review summaries. The assistant also handles post-purchase queries like "When will my order arrive?" by pulling live tracking data from the merchant’s fulfillment system. This is not just a search upgrade—it’s a full lifecycle engagement tool.

Walmart Sparky: AI Meets Live Shopping

Walmart’s Sparky AI assistant made its live shopping debut in June 2026 as part of the retailer’s Summer Deals campaign retailtechinnovationhub.com. Unlike static assistants, Sparky operates within live video streams—an increasingly popular format for e-commerce where hosts demonstrate products in real time. Viewers can ask Sparky questions such as "What’s the return policy on this blender?" or "Show me this in blue," and the AI responds instantly, surfacing relevant product pages and coupons.

This integration of AI with live shopping is noteworthy for several reasons. First, it bridges the gap between entertainment and utility: a shopper watching a host grill steaks can ask Sparky for the exact model number of the grill, its price, and shipping options without pausing the stream. Second, it reduces the need for human moderators to handle repetitive questions, allowing the stream’s host to focus on demonstration and engagement. However, early user feedback suggests that Sparky occasionally misinterprets requests or provides irrelevant answers, a reminder that even advanced AI assistants still have quality-control challenges.

What eBay Did Wrong: The Warning from Particle-MP

Not every AI shopping assistant is welcomed. A recent user report on EcommerceBytes details how eBay’s AI pop-ups are disrupting the shopping experience ecommercebytes.com. According to a user known as Particle-MP, eBay’s AI assistant generates intrusive pop-ups that obscure product listings, slow down page loads, and often suggest irrelevant items. The user wrote that the pop-ups "disrupt the shopping experience" and that the suggestions are "rarely helpful." This mirrors broader criticisms of AI deployed without adequate user testing or opt-in mechanisms.

The contrast with Goddiva AI and THG Ingenuity’s assistant is instructive. Goddiva’s assistant is embedded in a chat window that users can choose to open; it does not hijack the browsing flow. Walmart’s Sparky is contextually relevant to the live-stream environment. eBay’s implementation, by contrast, interrupts the core search-and-compare workflow.

Assistant	Integration Approach	User Sentiment
Goddiva AI	Chat widget (opt-in)	Generally positive
THG Ingenuity + Google Cloud	Embedded in search/support (passive)	N/A (enterprise)
Walmart Sparky	Live streaming overlay (semi-opt-in)	Mixed
eBay AI	Interstitial pop-up (forced)	Negative

For retailers considering an AI assistant, the eBay case offers a clear lesson: respect the user’s flow. An AI assistant that adds friction—rather than removing it—will repel customers, regardless of how advanced the underlying model is.

The Broader Implications for Retail AI

The three successful launches (Goddiva, THG Ingenuity, Walmart) and one misstep (eBay) illustrate several emerging principles for AI shopping assistants in 2026.

1. LLM-powered assistants are becoming table stakes. Two years ago, a chat-based shopping assistant was a novelty. Today, it’s an expected feature for competitive retailers. The Goddiva and THG launches show that brands are investing in custom-built, model-specific integrations rather than generic chatbot templates.

2. Timing matters. Walmart’s Sparky debuts alongside a major promotional event (Summer Deals), ensuring high visibility and a clear use case. Goddiva launched its assistant ahead of the holiday shopping season. eBay, by contrast, seems to have rolled out its AI pop-ups without tying them to a specific campaign or user benefit, which may explain the negative reception.

3. Personalization must be earned. Goddiva’s assistant offers personalized discounts and proactive suggestions—but only after the user initiates a conversation. That consent-based approach builds trust. eBay’s forced pop-ups feel like surveillance. Retailers that prioritize user agency will see better engagement and lower abandonment rates.

4. Live shopping is AI’s next frontier. Walmart’s Sparky experiment signals that AI assistants will increasingly complement, rather than replace, human hosts in live-stream commerce. The real-time Q&A capability solves a genuine pain point—viewers can’t pause to search while a host is demonstrating—and keeps attention on the stream.

What’s Next: From Chat to Autonomous Shopping Agents

Looking ahead, the capabilities demonstrated by Goddiva’s GPT-5 assistant and THG Ingenuity’s Google Cloud integration point toward a future where AI shopping assistants evolve from passive responders to proactive agents. Imagine an assistant that monitors a user’s calendar, detects an upcoming birthday, and preemptively suggests gift options—complete with price comparisons and discount codes—without any prompt. The technological pieces (LLMs, real-time data access, personalization engines) are already in place. The remaining challenge is user trust.

The Whatnot acquisition of Shaped techcrunch.com further underscores this trend. Shaped provides real-time recommendation infrastructure, which Whatnot will use to power AI-driven suggestions during live streams. This mirrors Walmart’s Sparky approach and suggests that real-time, context-aware recommendations will become a standard feature in live shopping platforms.

Final Takeaways

AI shopping assistants in 2026 are no longer a futuristic concept—they are live in production at retailers ranging from boutique fashion brands to global giants. Goddiva’s GPT-5-powered assistant shows that even mid-size retailers can deploy state-of-the-art AI. THG Ingenuity’s partnership with Google Cloud demonstrates that enterprise platforms are integrating AI at scale. Walmart’s Sparky proves that AI can enhance live shopping experiences. And eBay’s misstep serves as a cautionary tale about the importance of thoughtful user experience.

Retailers that deploy AI shopping assistants thoughtfully—with opt-in interactions, clear value, and respect for user flow—will gain a competitive edge. Those that force AI onto shoppers without consideration for context risk driving customers away.