Most conversations about AI shopping in 2026 collapse into two names, and treating those two as the whole map is how brands miss channels where the competition is thinner and the shoppers are just as ready to buy. ChatGPT and Gemini dominate the discourse because they are the largest and the loudest, and it is easy to conclude that agentic commerce is a two-horse race between OpenAI and Google. It is not. There is a second tier of assistant surfaces where people are actively shopping, where the major protocols already reach, and where far fewer brands have bothered to show up, which is precisely what makes them worth your attention. The store that broadens its view beyond the two giants finds demand that its competitors, fixated on the headline names, are not competing for.

Consider the surfaces that the coverage tends to skip. Perplexity built its reputation as an answer engine and extended into shopping, letting users research and, for its subscribers, purchase products within the experience. Microsoft's Copilot is embedded across an enormous installed base of productivity software and the Windows environment, and Microsoft has been building commerce capabilities into it, giving it a shopping surface reaching people in the middle of their working day rather than only in a dedicated shopping session. Klarna, known for payments, operates an AI assistant that helps people shop, sitting at the intersection of discovery and financing where purchase intent is unusually high. None of these is ChatGPT-sized, but each reaches a real and distinct population of shoppers, and each is a place a recommendation can happen.

The reason this matters practically is that these surfaces are not separate universes requiring separate feeds and separate integrations from scratch. The open protocols were designed to reach across the ecosystem, and support for the standards that carry your catalog into agentic answers has been confirmed across surfaces including Microsoft Copilot, Klarna's assistant, and Perplexity, alongside the Google surfaces. Shopify's move to make its merchants discoverable did not stop at one assistant either. It positioned stores to be sellable across ChatGPT, Microsoft Copilot, and Google's surfaces from a single setup. This is the crucial efficiency. The clean, complete, structured catalog data you build to win in ChatGPT and Gemini is largely the same data that makes you legible on these other surfaces. You are not doing the work three or four times. You are doing it once and reaching further, provided you actually understand that the further reach exists and is worth claiming.

There is a competitive logic here that mirrors the advantage of any underattended channel. When every brand piles into optimizing for the two largest assistants, the field on those surfaces gets crowded, and standing out gets harder. On the second-tier surfaces, where fewer brands have made themselves visible, the field is thinner, and the same clean data that merely keeps you competitive on ChatGPT can make you conspicuously present on Perplexity or inside Copilot. This is the classic pattern of early channels. The advantage goes not to the brand that optimizes hardest for the crowded surface but to the one that shows up on the surfaces its competitors have not yet noticed. Being one of the few legible brands in a shopper's Perplexity or Copilot answer is worth more, per unit of effort, than being one of many in their ChatGPT answer.

It also pays to think about who these surfaces reach, because the audiences are not identical and the differences are strategically useful. Perplexity attracts people in a research posture, comparing and investigating before deciding, which means a brand that is legible there catches shoppers at the point where a shortlist is forming rather than after it has hardened. Copilot reaches people inside their work environment, where a shopping need might surface mid-task and be acted on immediately, a context that ChatGPT's dedicated sessions do not fully capture. A financing-adjacent assistant reaches shoppers at the moment they are thinking about how to pay, which correlates with high and immediate purchase intent. Appearing across these surfaces is not just about reaching more people. It is about reaching people in different moments of the buying journey, each of which a single-surface strategy misses.

The skeptical caveat is worth stating so the enthusiasm stays grounded. These surfaces are smaller, some of their commerce capabilities are newer and still maturing, and the volume any one of them sends you today may be modest compared to the giants. Treating them as your primary channel would be a mistake in the opposite direction, over-rotating onto small surfaces at the expense of the large ones. The correct posture is not to abandon ChatGPT and Gemini for the second tier. It is to make sure that the data work you are already doing for the giants extends to the second tier as well, so you capture the incremental demand there at nearly zero marginal cost, while keeping your center of gravity on the largest surfaces. This is a both-and strategy, not an either-or, and its efficiency comes from the fact that the underlying data is shared.

There is a durability argument too, which is that the assistant landscape is not settled and betting everything on today's two leaders is a concentration risk. The relative standing of these surfaces will shift, new entrants will appear, and the population of shoppers will distribute itself across them in ways no one can precisely predict. A brand whose visibility strategy is built around clean, protocol-legible catalog data is resilient to that churn, because that data serves whichever surfaces rise, while a brand that hand-optimized narrowly for one or two assistants is exposed if the distribution moves. Building for the ecosystem rather than for two specific products is how you avoid having to redo the work every time the competitive order rearranges itself, and the second-tier surfaces are a useful reminder that the ecosystem is broader than the headlines suggest.

For the store owner who wants the practical takeaway, it is straightforward and reassuringly cheap. Do not let the two-name framing define your channel strategy, because it leaves real demand uncaptured on surfaces your competitors are ignoring. Confirm that your catalog is legible not only to ChatGPT and Gemini but to the second-tier assistants where the protocols already reach, including Perplexity, Copilot, and financing-adjacent shopping assistants, recognizing that on many platforms this extended reach comes from the same setup. Lean on the fact that the clean, structured data you build once serves all of these surfaces, so the incremental effort to appear on them is small. Keep your primary focus on the largest surfaces while treating the second tier as low-cost incremental reach into different moments of the buying journey. And build for the ecosystem rather than for two products, so your visibility survives the inevitable reshuffling of which assistant leads. The brands that win the next phase of AI shopping are not only the ones that optimized hardest for the giants. They are the ones that noticed the shopping was happening in more places than the headlines admitted, and showed up where their competitors did not think to look.

It is worth being concrete about how a store actually checks its presence on these second-tier surfaces, because the abstraction of reaching them means little without verification. The same manual discipline that reveals your visibility on the giants applies here, asking each of these assistants the questions your shoppers ask, in your categories, and observing whether and how you appear. This takes little more than the willingness to use the surfaces yourself and record what you find, and it converts a vague hope that your data extends to them into a concrete picture of whether it actually does. A store that assumes its clean catalog automatically wins everywhere, without ever checking the second-tier surfaces directly, is trusting rather than verifying, and the gaps that trust hides are exactly the ones that quietly cost recommendations. Checking is cheap. Assuming is where visibility silently leaks.

The deeper strategic value of these surfaces is that they hedge against a future no one can predict, because the distribution of shoppers across assistants is genuinely unsettled. If a second-tier surface grows faster than expected, a brand already legible there rides that growth rather than scrambling to catch up, while a brand that ignored it starts from behind. Because the cost of extending to these surfaces is so low when the underlying data is shared, treating them as cheap insurance against the reordering of the assistant landscape is simply prudent. You are not betting on any one of them to win. You are ensuring that whichever surfaces rise, your products are already present on them, which is a far more comfortable position than having concentrated all your visibility on the two names that happen to lead today.

The practical instruction, then, is to widen your definition of the channel, verify your presence on the surfaces beyond the giants rather than assuming it, and treat the low marginal cost of that reach as a reason to claim it rather than an excuse to defer it. The shopping is happening in more places than the two names everyone repeats, and the brands that show up where their competitors are not looking are the ones that quietly capture the demand those competitors never see.