In this episode of Talk Commerce recorded live from Ecom Forum in Minneapolis, host Brent Peterson sits down with Sharon Gee, Senior Vice President of Product at Commerce, to discuss the seismic shifts happening in ecommerce. The conversation explores how artificial intelligence and agentic commerce are reshaping the way merchants connect with customers. Sharon brings extensive experience from her six years at Commerce, where she oversees AI offerings across BigCommerce, Feedonomics, and Makeswift. What emerges from this discussion isn’t just another tech conversation but rather a roadmap for merchants navigating the transition from traditional SEO to a world where agents shop alongside humans.
Key Takeaways
- Data has become the new storefront as consumers increasingly turn to answer engines rather than traditional search
- Merchants need to provide structured, contextual data to AI agents, not just visually appealing websites for human shoppers
- The adoption rate of AI tools like ChatGPT has outpaced every other consumer technology in history, including cell phones
- Product data must now exist on multiple levels, from basic ad information to unstructured content hidden in PDFs
- B2B commerce stands to benefit significantly from agentic AI, particularly through AI-powered sales assistants
- Trust protocols are being established to manage transactions between shoppers, shopper agents, merchants, and merchant agents
- AI democratizes marketing tools, allowing creative thinkers to execute ideas without engineering expertise
- User reviews represent a treasure trove of search terms that should inform product descriptions
About Sharon Gee
Sharon serves as Senior Vice President of the Product Organization at Commerce, where she focuses on AI offerings across the company’s portfolio. She played a key role in leading the acquisition strategy for Feedonomics four years ago and served as General Manager of that business during its successful integration. Before joining Commerce, Sharon spent time agency-side in New York City. Her expertise spans ecommerce platforms, enterprise data feed management, and visual editing solutions. Outside her professional life, Sharon owns a flower farm and coffee shop in Colorado, offering her a unique perspective that balances digital commerce with hands-on retail experience. Throughout the industry, Sharon has become recognized for her insights on how AI and data optimization can transform merchant visibility and customer acquisition.
Episode Summary
The conversation begins with Sharon outlining her role at Commerce and immediately diving into what she describes as the most exciting development in ecommerce: agentic commerce. She explains that for decades, commerce professionals have been optimizing data for advertising channels, trying to improve conversion rates and return on ad spend. However, the fundamental rules remained consistent—acquire customers through Google or Meta, drive them to your website, and hope to convert them at rates between two and five percent.
“Somebody came along and bopped the board game and now we get to reset all the pieces,” Sharon explains. The game-changer is that consumers now turn to answer engines for their most basic questions. These aren’t simple queries based on price or size filters. Instead, shoppers ask complex questions like wanting a dress for a wedding in Italy in a specific color and size, delivered by tomorrow. This shift requires merchants to bring together data from marketing channels, internal systems, and content teams because data has become the new storefront.
Sharon emphasizes that answer engines need deep context to respond to long-form queries effectively. The challenge for merchants becomes ensuring their products are discoverable wherever shoppers are looking and making it easy to shop however the consumer prefers—whether that means clicking through to a personalized product page where they can visualize furniture in their living room or buying mascara with a thumbprint because they already know what they want.
The conversation then shifts to the technical differences between old SEO practices and the new reality of AI-driven discovery. Sharon points out that ChatGPT reached 100 million users faster than any other technology in history. This rapid adoption creates both opportunity and challenge. Answer engines need data, and while they can scrape websites for it, those websites aren’t optimized for agents. They’re full of HTML, images, and visual elements designed for human brains, not for AI consumption.
Sharon introduces a framework for thinking about data levels. Level one includes basic information needed for Google ads—title, description, image, size, color, and weight. Level two encompasses the significantly more extensive data required to list on marketplaces like Amazon. Level three consists of product specifications sitting in Product Information Management systems—manufacturing details, materials, origins, and technical specs. Levels four and five venture into unstructured data territory, including PDFs on websites and user reviews.
“That’s not the kind of data you usually show on a product detail page,” Sharon notes. This creates what she calls a bifurcated experience. Merchants now need to provision different experiences because agents are customers too. When an agent visits a site, it doesn’t need pretty pictures—it needs structured data and links to images it might want to reference.
Brent raises the question of whether this means adding data below the fold on product pages or creating entirely separate experiences. Sharon confirms the latter. When a merchant senses that an agent rather than a human is visiting, they should render a different version of the website filled with data rather than images. This aligns with what Sharon identifies as three fundamental truths: the customer is the channel, data is the new storefront, and agents are customers too.
The discussion moves to whether merchant sites might eventually become pure APIs without customer-facing elements. Sharon argues for a both-and approach. The brand site remains one channel where people interact with data, and it’s the one channel merchants fully control. However, on third-party agentic channels, merchants don’t control visualization—they only control the data they provide. This makes data investment critical for visibility on channels merchants don’t control, while simultaneously requiring deep investment in owned channels.
Sharon draws a parallel to how marketers have always known that sending better data to Google results in lower cost-per-click because the data more relevantly answers searcher queries. She observes that data specialists are inheriting the earth—the people who once led organic search, then paid advertising, now lead agentic strategy. This mirrors how creative directors once ran websites before being replaced by people who could read website analytics.
The conversation touches on both first-party and third-party AI applications. Sharon describes the baby version of what’s coming as shopper assistants or chatbot experiences on brand websites. However, she sees massive potential in B2B sales assistants trained on the same documentation as human sales representatives. If three-quarters of the sales cycle could progress overnight while sales reps sleep, those reps could focus on high-touch human interactions. Sharon believes B2B commerce will leapfrog B2C experiences through agentic AI because B2B companies are manufacturers with deep data, extensive documentation, and sophisticated pricing structures with custom price books and customer groups.
Brent raises concerns about AI reliability, noting his frustrations with coding assistants that make illogical mistakes and assumptions. He envisions scenarios where an agent searching for hiking shoes for Tuscany presents three options but autonomously purchases one without confirmation. Sharon acknowledges these valid concerns and explains that commerce platforms, channel partners, and payment partners are collaborating on protocols to address exactly these issues.
“You’ve seen more open protocols released in the past six months than like the previous 10 years combined,” Sharon observes. Companies across the industry recognize that nobody wants an internet that isn’t safe or trustworthy. Trust becomes paramount when authorizing agents to shop on behalf of consumers. The human-in-the-loop component requires careful protocol design because transactions now involve four parties: a shopper, a shopper agent, a merchant, and a merchant agent. All four must trust each other.
Sharon mentions specific initiatives like Stripe ACP and PayPal protocols, as well as Google’s AP2 and other agentic protocols. Technology companies are leaning into these challenges because the problems are both complex and exciting. Meanwhile, attorneys are appropriately concerned about data security. Sharon frames this moment as one where the new rules of the internet are being written in the agentic space.
The opportunities this creates excite Sharon tremendously. She asks Brent to imagine rewriting an entire product catalog with a button click using generative AI, based on search terms from various channels. A merchant could refocus their entire catalog around Halloween instantly. Previous limitations—insufficient copywriters or creative resources—no longer apply. While many discuss AI primarily as a cost-reduction tool for operational efficiency, Sharon emphasizes its role as a growth enabler. AI provides jet fuel for existing team members, unlocking capabilities and scale never before possible because humans are freed from operational tasks that robots handle better.
“I would love it if our generation is the last one to use a mouse and a keyboard,” Sharon declares, capturing her optimism about AI’s potential to improve user experiences fundamentally.
Brent agrees and adds that AI’s greatest value for merchants might be identifying what they’re not doing rather than what they should be doing. Instead of worrying about generating content, merchants should focus on finding patterns in their data that reveal missing content opportunities.
Sharon confirms that many Commerce customers use tools to define simulated personas based on actual users, then understand what queries those personas might ask on various channels. Based on those questions, merchants can determine what content they need. She returns to the example of someone in Colorado planning an Italy vacation—does a merchant have the right content to ensure their products get referenced instead of competitors’ products?
Sharon believes marketers who understand what shoppers actually want and can articulate their unique value proposition will win because AI has democratized tooling. All platforms are working to ensure an open, trusted transactional experience with secure data presentation. For brand marketers, this represents an extraordinary opportunity. An army of agents can now support goals that previously required engineering expertise. If someone can think it, dream it, and believe it would deliver good outcomes, they can do it.
As the conversation concludes, Sharon reflects on why she values Ecom Forum. She praises Darin and the Titans group as heartfelt humans in commerce who curate thought leaders dealing with real implementation problems. Despite AI’s omnipresence, Sharon reminds listeners that commerce still centers on humans. Sharon Gee’s insights reveal that success in this new landscape requires merchants to embrace data as their most valuable asset while never losing sight of the human experiences they’re ultimately trying to enhance.
Final Thoughts
The transformation Sharon describes isn’t coming—it’s already here. Merchants who recognize that data has become their new storefront and invest accordingly will capture outsized visibility in channels where attention is rapidly shifting. The bifurcation between human and agent experiences requires technical sophistication, but platforms are building the infrastructure to make this transition manageable. What remains constant is the need to understand customers deeply and articulate unique value clearly. As protocols establish trust frameworks for this four-party transaction ecosystem, the merchants who win won’t just be the ones with the best technology. They’ll be the ones who recognize that while agents are shopping, humans are still the ones making the final decisions—and both deserve experiences built specifically for them. In the end, you might say the future of commerce isn’t just about making transactions easier—it’s about making discovery more intelligent and trust more transparent, one data point at a time.
Follow Talk Commerce on your favorite platform:
