# Vinod Kumar

Vinod Kumar writes about how AI is changing ecommerce discovery and merchandising. Kumar's work includes auditing how AI tools surface brands and examining how shopper agents should guide product discovery.

- Profile: https://b2binfluencer.marketing/creators/vinod-kumar/
- LinkedIn: https://www.linkedin.com/in/vinodkumar-ai
- LinkedIn followers: 3,730
- Topics: artificial-intelligence, ecommerce, agentic-commerce, answer-engine-optimization, product-discovery
- Based on 15 dated posts, latest 2026-10-07

## Summary

They are most useful for explaining AI ecommerce discovery and merchandising through storefront audits and shopper agent examples.

- Best for: A B2B ecommerce software webinar or content collaboration for merchandising and digital product teams.
- Not for: A consumer lifestyle brand seeking broad awareness or verified demographics and guaranteed distribution.
- Check first: Confirm deliverables, review terms, audience role, campaign reach, and comparable engagement results.

### Takeaways

- PUMA appeared consistently in ChatGPT but was absent from Gemini in 29 of 36 shopping queries.
- The Frankenstack tax is the coordination cost when separate systems make competing recommendations for one shopper.
- YETI reports a 2X conversion lift for Ranger.

## Main ideas

- **Conversational shopping is a merchandising problem, not a language-model problem**: Vinod Kumar argues that answering a shopper's question and merchandising a store are different jobs that vendors constantly conflate. They break product discovery into five functions: resolving a specific fact such as sizing or a return policy; retrieving products by matching a query to a catalog; reading the intent behind a phrase; deciding what gets shown; and applying the merchant's accumulated judgment about what belongs in front of a customer and why. Most AI shopping vendors have built the first two, and marketing often presents the third as the finished product. The fourth and fifth are ranking and curation choices shaped by margin, inventory position and what the brand wants to be known for—not language processing. Their summary: the chatbot problem is largely solved, while the merchandising problem remains open and mostly unnamed.
- **The Frankenstack tax: the cost sits in the seams between tools, not in the tools themselves**: Vinod Kumar audits live storefronts and counts the systems independently deciding what a shopper sees. Recommendation-adjacent work is repeatedly split across vendors that are not required to agree, so the same shopper can get different "relevant" products on the cart page and the product page during one visit. The expense lies not in buying the tools but in stitching shopper context back together: engineering hours, conflicting signals and an inconsistent experience depending on which system responds. They argue this is structural, not a platform flaw, and that fragmented context—not model capability—will hold agentic systems back.
- **AI visibility is plural—a single score hides which assistant is failing and why**: Vinod Kumar argues that brands should stop treating AI visibility as one number. A brand can look healthy in ChatGPT and be nearly absent in Gemini, while an aggregate score conceals the actual merchandising gap. They split the problem into two questions: are you mentioned at all, and what does the AI say about you when it finds you? Those are different optimization tasks, but vendors tend to merge them. They also argue that answer-engine optimization is not search-engine optimization in a new bottle: some techniques carry over, but ranking well in keyword search says little about being retrieved by an AI assistant, and the brand controls none of those new shelves.
- **Discovery increasingly skips the storefront, so every product page becomes a landing page**: Vinod Kumar uses the term D2P, Direct-to-PDP, for shoppers who start in ChatGPT, Gemini, Perplexity, TikTok, Instagram, Reddit, YouTube or AI search, ask a specific question, get a curated recommendation and land directly on a product detail page, bypassing the homepage, category page and search results page. Product data, structured content, reviews, FAQs, availability and brand context become the raw material AI systems use to decide what gets recommended. If a product can be bought without first visiting the storefront, merchandising strategy cannot stop there.
- **Trust sets the pace of agentic commerce, not convenience**: Vinod Kumar argues that agentic commerce frames shopping as a friction problem to remove—find, compare, buy faster—and assumes convenience will drive adoption. History suggests otherwise: ecommerce took years to build trust around payments, marketplaces took time to build trust between strangers, and digital wallets became routine only gradually. Trust builds more slowly than technology. Consumers may be comfortable using AI to retrieve information, but delegating an actual buying decision is a different threshold and will take longer to cross. This caution sits alongside their enthusiasm for branded shopper agents, sharpening rather than weakening the point: agents can work today while mass delegation of decisions remains slow.
- **Relationships are treated as a working method, not a soft topic**: Alongside the audits, Vinod Kumar states a simpler priority: relationships drive the industry, and founders should lean on existing ones to find traction. They describe sitting down with the Rewind team at Shoptalk Vegas to discuss the future of AI product discovery, bringing their own startup experience into the conversation. It is the shortest and least technical of the posts, but shows how they connect event conversations, peer relationships and go-to-market.

## Work to read

- [The PUMA AI visibility audit: ChatGPT knows the brand, Gemini does not](https://www.linkedin.com/posts/vinodkumar-ai_if-you-run-dtc-for-a-lifestyle-brand-and-activity-7500165375595200512-eKYA)
- [Five functions of product discovery, learned launching YETI's assistant Ranger](https://www.linkedin.com/posts/vinodkumar-ai_for-brands-rushing-to-launch-a-shopper-agent-activity-7498370757123772417-Z7kE)
- [YETI reports a 2X conversion lift for Ranger](https://www.linkedin.com/posts/vinodkumar-ai_when-a-consumer-engages-with-ranger-we-see-activity-7506698855288086528-Yk0X)
- [The Frankenstack tax, part one: 15 vendors on Under Armour's storefront](https://www.linkedin.com/posts/vinodkumar-ai_the-biggest-issue-with-adopting-agentic-systems-activity-7475881305168093184-r_5f)
- [Frankenstack follow-up: Arhaus shows the same pattern on Shopify Plus](https://www.linkedin.com/posts/vinodkumar-ai_after-i-published-the-under-armour-storefront-activity-7476268258589859840-32GG)
- ["D2P Shopper Journeys": discovery that skips the storefront](https://www.linkedin.com/posts/vinodkumar-ai_i-came-across-a-new-term-today-d2p-shopper-activity-7470112485073694720-gvMO)
- [Why delegating purchases to AI will take longer than the technology cycle](https://www.linkedin.com/posts/vinodkumar-ai_the-ongoing-discussion-about-agentic-commerce-activity-7475148012076343296-FgR3)
- [Advice from Shoptalk: lean into existing relationships](https://www.linkedin.com/posts/vinodkumar-ai_relationships-drive-this-industry-at-shoptalk-activity-7465008841768972288-NDHN)

## Brand relationships

- **Syntheum AI** (owned_business)
- **YETI** (client)
- **PUMA** (mention)
- **Under Armour** (mention)
- **Arhaus** (mention)
- **Rewind** (mention)

## Partnerships

The material reviewed contains no published partnership offer for Vinod Kumar: no sponsorship menu, media kit, rate card, or stated terms for a webinar or paid collaboration. Their visible work includes storefront and AI-visibility audits, links in post comments to longer articles and an audit agent, and posts as part of Syntheum AI about the team's work. The clearest example is YETI's conversational shopping assistant, Ranger, which they describe as a Syntheum AI project and credit to YETI's digital products and merchandising team, citing a reported 2x conversion lift. The public call to action is a product audit for a brand's own use, not a partnership inquiry route. Buyers should open a direct conversation and confirm the deliverables, whether the brand can review the analysis before publication, how brand names, storefront data, and audit findings may be used, and who owns the resulting material.

## Alternatives

- Max Sinclair: Choose Max Sinclair when the session needs to be current and news-driven rather than built around a tested operating framework. Their posts move fast on agentic-commerce announcements: one covers OpenAI hiring Peter Steinberger, the creator of the open-source agent OpenClaw, to build personal agents, and draws out what agent-handled routine purchases mean for consumer brands. Another covers Walmart tightening product titles to 90 characters so its Sparky shopping agent can parse listings, citing quarterly figures for Sparky users and order value. For a B2B ecommerce software company running a webinar or content collaboration on AI visibility and shopper agents, this is useful when the promise is “what just changed and what to do about it,” with fresh marketplace and platform examples.
- Aabhas Vijay: Choose Aabhas Vijay when marketers, SEO leads, or agencies need a repeatable method for getting cited by AI assistants, rather than a merchandising view of what an assistant should recommend. Their posts offer a working procedure: don't start from keywords; reverse the usual order; test whether ChatGPT actually performs web retrieval for the category; map the query fan-outs behind a single prompt (they cite an average of 7.6 fan-out searches per prompt); and study which sources get cited, including data suggesting a large share of citations go to low-authority domains. Worked examples include a CRM query built from a single buyer question. For a campaign aimed at brand or agency teams, this suits a session where attendees want to leave with a method they can run next week.
- Tom Shivers: Choose Tom Shivers when ecommerce operations or technical teams need the plumbing that determines whether an agent can recommend or transact at all. Their posts explain OpenAI and Stripe's Agentic Commerce Protocol and Google's Universal Commerce Protocol in plain terms—what each lets an agent compare, build carts, and check out without visiting a retailer's site while the retailer stays merchant of record. They also offer concrete steps: structured product feeds, Merchant Center Next, schema and brand-entity work, delegated payments, and fulfillment and policy data that machines can read. One post offers a feed-readiness checklist for auditing brands. For a campaign aimed at teams considering a session on agentic commerce, this pairs well with Vinod Kumar's merchandising framing: one side covers what the agent needs technically, the other what a merchandiser should decide. A joint session could move from feed readiness to recommendation choice.

---
AI-assisted summary of public posts. Audience demographics are not available. Engagement appears only where we checked recent posts. Method: https://b2binfluencer.marketing/methodology/