Who doesn’t love fall? That first touch of crisp air, pumpkin getting shoehorned into every beverage and dessert, and, if you’re in marketing for retail, CPG, or travel brands, furiously preparing for the holiday shopping season.
This holiday season is going to feel wildly different for both marketers and shoppers. Merkle’s September 2026 research found more than half of shoppers now turn to AI for gift ideas. They’re also leaning on it for brand and product recommendations and even the full purchase process. That’s more ways than ever to end up on the naughty list.
2026 is officially the first AI-assisted holiday shopping season where your customers can complete their entire research and purchase journey inside of an LLM. There’s a lot of promise and opportunity: agentic commerce, maturing toolsets, and new features inside answer engines and traditional search engines are moving from potential to reality. But as always, with opportunity comes pressure to have a well-laid-out strategy for optimization.
As we all navigate this rapidly evolving world where customers outsource their shopping to their favorite AI platform, let’s look at some ways we expect brands to get caught with empty stockings.
Prediction 1: A challenger brand will outperform an established category leader
Actually, it’s probably more like hundreds of challenger brands, not just a single one.
AI answer engines and shopping agents prioritize whichever brand gives them the clearest, most complete, confidently structured answer to the question being asked. A category leader with a massive catalog, inconsistent product data, and years of technical SEO debt gives the system less to work with than a smaller competitor that’s invested in getting the basics right: clean structured data, clear entity signals, and content that actually answers the question instead of talking around it.
That means doing the unglamorous work: fixing the schema, writing product descriptions like someone asks a real question, and keeping things current instead of setting it and forgetting it.
Prediction 2: Partial optimization will become full-blown failure
Someone’s IT team is going to lock down bot traffic ahead of Black Friday to make sure they don’t get knocked offline during the biggest week of the year. But the rule set they write to do it will block the exact AI crawlers they need to read their site. Nobody will notice until the numbers come in soft.
Somebody else will earn a spot in a dozen “best gifts for” roundups, except half the guide’s details trace back to an unmoderated Reddit post containing information that was never true, written and indexed weeks before the season even started. What looks like hard-won credibility is actually misleading potential customers.
Somebody else will have a rushed seasonal landing page, built quickly under a strict deadline. It has fantastic content, properly structured, but it’s invisible to some AI crawlers because of how it’s rendered.
We could keep going, but you get the idea. You can ace three-quarters of the test and still fail it because AI systems don’t grade on a curve. They pull from everything they can find about you, not just the part you happened to invest in this year.
Prediction 3: Customer support will field complaints about bad AI intel
A version of this is already happening. Someone calls your 800 number, navigates through your phone tree, and finally reaches a rep. They’re confused. They asked an AI assistant to compare noise-cancelling headphones for a holiday gift. The assistant told them your model was discontinued last spring and recommended a competitor instead. Except your model wasn’t discontinued; it’s sitting on shelves, ready to ship. The AI picked that up from an eight-month-old forum thread where someone speculated your company was “probably killing off the line.” Nobody ever corrected it, so the guess calcified into something the model now states as fact.
Nobody on your team said anything wrong. Your own site is accurate. The AI just trusted an old, stale source somewhere else on the internet more than it trusted you, and you had no way of knowing that source existed until a customer showed up confused because of it.
Prediction 4: Brands will feel the limits of their own measurement
Many brands have invested in a monitoring tool to track AI visibility. They can see when they show up in AI answers and when they don’t. They can benchmark against competitors and start to understand where they show up. The problem is, showing up isn’t the same as showing up correctly. Well-rounded measurement goes beyond the where and when to answer the what and why.
Even a brand that catches an issue has to do something about it. That means testing a fix, watching whether it actually worked, and doing it fast enough to try again before the season ends. In our experience, most brands don’t have that in place yet. They notice something’s off, make a change, and hope.
Are you visible? Is what AI says about you correct? Even brands that can answer yes to both will end up on the outside looking in this season, because there’s no time left to test optimization strategies and prove they worked before the holidays are over.
Prediction 5: Web traffic will skip the journey and go right to sales
You’ve spent enormous effort building out your digital customer journeys, presenting relevant, empathetic copy on your top-funnel pages, and delivering personalized experiences that recommend the right thing based on behavioral and data signals. So why is everyone skipping all of that and landing straight on a product page, ready to buy?
This isn’t the direct-to-PDP traffic you already know how to read. A shopper who clicks a product ad chose your brand’s paid placement, and you can trace the channel, measure the return, and retarget them if they leave. A shopper who arrives after asking an AI to compare options looks the same in your analytics, with little to no context on what they searched or their intent—just that they were referred from Google or ChatGPT. They already ran the comparison, read the reviews, and picked a winner somewhere you can’t see and can’t correct.
During the holiday stretch, when gift deadlines and shipping cutoffs compress the whole decision into one sitting, more of that comparison shopping happens off your site. Marketing teams will watch traffic drop on nearly every top-funnel page while product pages spike, and the conversion rates on those pages will look better than ever. That looks like a win until you notice what’s missing: no way to tell how many of those conversions started with an AI recommendation, no way to fix it if the AI got something wrong, and no way to know if you’ll get picked again.
You weren’t there for the moment someone got excited about buying from you. An AI was. If the numbers are there, what’s the problem? Customer lifetime value runs on trust, and nothing in a conversion report tells you whether that trust belongs to you or to whatever answered the question first.
Where this leaves us
Even if these challenges sound familiar, you shouldn’t panic or scrap what you’ve built. You’re already a step ahead: you’re aware of what many brands will find out the hard way during the most critical weeks of the year. Focus on checking your data, your coverage, and whether your measurement tells you what’s both visible and true.
Living through Q4 will expose exactly where you need to go next, giving you clarity to map your AI discoverability gaps for Q1 with confidence. Do that, and you land on the nice list. If you need help prioritizing those gaps, we’re always here to talk.
