This story was first reported on and published by Glossy sibling site Modern Retail.
These days, shoppers are increasingly asking ChatGPT for recommendations like “the best sleep supplement” and “best women’s multivitamin.” So instead of browsing a brand’s homepage, shoppers are landing directly on a product page. And they may be deciding whether or not to purchase within minutes.
As more research and product discovery happen through LLMs and AI platforms, brands like supplement maker Olly say they are experiencing a shift in their traffic patterns in real time. According to Jennifer Peters, Olly’s director of DTC, marketing technology and digital compliance, shoppers are now coming to the brand’s website from a lower part of the marketing funnel. Due to the high intent of purchase, these shoppers are often ready to check out much quicker than other website visitors. As such, she said, visitors are spending less time clicking around the Olly site, but are converting faster to the products.
This shifting behavior is forcing brands to focus more on the actual product pages and less so on the home landing pages. Olly is making changes to its website by beefing up its in-house educational blog posts and expanding each product page with a frequently asked questions section.
Peter said the majority of Olly’s revenue comes from in-store sales at retailers like Walmart and Target. “But our DTC site will always be a small but necessary piece of revenue when looking at the overall pie,” she said.
But what has changed and is continuing to evolve, Peters said, is that suddenly a brand’s DTC website is the point of entry for discoverability as people are using LLMs to ask questions or find health solutions. “People are now entering the mid-funnel instead of top-of-funnel, which is what we’ve been doing the last 15 years,” she said.
Currently, the majority of Olly’s traffic is still coming from social media and digital ads. But Peters anticipates traffic from LLMs will increase with time. “We sell to millennial women, and we always have,” Peters said. “So as millennial women adapt to this new way of search, we’ll see more of it.”
Peters said that, to address this newfound customer journey, the brand has two buckets of content to bring up to date when it comes to LLM discoverability. There is the internal content on the website, and content through third-party avenues like Reddit or social media influencers. “If a piece of information is inaccurate or out of date, that is going to hurt how your brand shows up,” Peters said.
“One of the first things we did was rewrite our Wikipedia page,” Peters said. “No one had touched it in about eight years, so it was outdated and old.” The new entries have yet to be published because Wikipedia must vet and verify the information when companies update their own pages.
To adapt to the new traffic source on its website, Peters said Olly used digital analytics company Contentsquare to identify friction — such as vague ingredient and solution descriptions — and optimize its pages for ChatGPT and Gemini. “We are most focused on what we own and what lives on our domain,” Peters said.
That led Olly to create an editorial calendar of education-focused articles and thought leadership pieces focused on the supplement space. This practice has helped other brands like Viv and Joe & Bella show up in AI search results due to their authoritative voice on specific topics.
Another major change Olly is beginning to implement this year is adding detailed FAQ sections to its product detail pages (PDPs), to better explain the ingredients and their efficacy.
“People are coming in saying, ‘I’m having a hard time sleeping. What should I do?’” Peters said.
The FAQ section fleshes out the ingredients and name checks supplements in the product, such as adaptogens, zinc or letter vitamins. Peters said this helps lead LLM users inquiring about sleep supplements straight to Olly’s Sleep gummies.
“Just having that information living on that page somewhere in an ‘FAQ’ format is something that LLMs like,” Peters said.
In a separate effort, the company is also updating landing pages to be clearer to the casual visitor, especially during Olly’s big annual sales pegged to New Year’s and back-to-school. This includes surfacing recommendations for high-performing products early on in the visitor’s browsing journey and developing a clearer “add to cart” call to action, she added. Olly also tweaked the language to clarify subscription options.
Peters said the company found that new customers wanted more clarity on the subscribe-and-save discount percentage they would receive. “Because if you’re coming to our DTC site to shop, you’re either a super fan of the brand or interested in subscribing and never having to worry about running out,” she said.
Those new checkout callouts on the product detail pages paid off. They resulted in a 20% month-over-month increase in website revenue earlier this year, along with a lift in Olly’s average order value.
Moving forward, brands have to build content for both human and AI discovery as LLMs continue to scale up.
Loni Stark, vp of strategy and product at Adobe, said more consumers are trusting AI chatbots as their primary channel for brand discovery, product evaluation and purchase decision making.
Recent Adobe data shows that AI-driven retail traffic increased by 393% year over year in the first quarter of 2026. These AI-prompted visits also demonstrate higher conversion rates, with retail AI-driven revenue per visit 37% higher than for non-AI traffic. “Just 12 months ago, non-AI visits were worth 128% more than AI visits, a seismic shift in how shoppers are making decisions,” Stark said.
This means it has become imperative for businesses to optimize every facet of their online content, for humans and AI agents alike. “If product information is outdated, or not even picked up by LLMs at all, a brand has a visibility gap,” Starik said. That could mean missing an opportunity to engage with shoppers who are increasingly relying on AI results when shopping online.
While traffic from AI results is increasing incrementally, Peters said Olly is not interested in producing generative AI content to chase chatbot traffic. “For example, we are not writing personalized content for anyone on the site,” she said. “That is not very scalable right now, but also we don’t want to publish fake content.”
Peters said that while AI search is yet another challenge for e-commerce brands, it also makes for an exciting time for growing DTC channels. “Because suddenly it’s not just about conversion, it’s about discoverability,” she said.
