YouTube has spent nearly two decades quietly assembling the largest library of product reviews in human history. Unboxings. Six-month check-ins. Side-by-side comparisons filmed in someone's garage under the lighting of a hostage video. For pretty much every product sold on Amazon or Shopify, somebody, somewhere, has filmed themselves holding it and talking about it for eleven minutes (plus a four-minute sponsor read for a VPN).

And for most of that time, nobody could find any of it.

That's the backdrop to YouTube rolling out Ask YouTube, its AI conversational search, to every US user across desktop, mobile and connected TV. Roughly half of American adults now use AI chatbots, so having one is less a strategy and more a dress code. But for those of us selling things online, the interesting story is smaller, nerdier and considerably more expensive: a conversational layer sitting on top of YouTube's archive changes which videos get found. And a remarkable number of those videos are product recommendations.

Keyword search matches what you type against a video's title, description and tags, then ranks results by relevance, engagement and recency. Brilliant if you already know the product name. Useless if you only know your problem.

Type the exact model of a blender and you'll find reviews of that blender. Type "a quiet blender for a small flat that can handle frozen fruit and costs under a hundred dollars" and the algorithm stares back at you like a waiter asked about the wine list in a language he doesn't speak.

The result is a long tail of genuinely useful content that never surfaces. A meticulous 2023 review from a small creator who tests kitchen appliances with the rigour of a forensic lab, sitting at a few thousand views, is effectively invisible. Meanwhile, the same handful of enormous channels win every product query, mostly by virtue of having already won every product query. (Search rankings and secondary school popularity contests: separated at birth.)

Conversational search works differently. Instead of matching words, a language model interprets intent, constraints and all, and retrieves content based on meaning. You can also refine without starting over: "what about one that's easier to clean?" No new search, no new tab, no new existential crisis.

YouTube hasn't been subtle about where this is heading. At Made on YouTube in September, it announced shopping features: product searches may return review videos organised into a comparison table built around the attributes you care about, and you can then ask follow-up questions right on a review's watch page.

People were using it well before the wide rollout: more than 140 million of them in June alone, up over 500% since the previous December. (So, not exactly a soft launch to friends and family.)

Here's the bit that matters for us. Meaning-based retrieval cares far less about a video's title, age or view count. In principle, it can recognise that a two-year-old review from a tiny channel answers a specific question better than last week's sponsored masterpiece from a creator with a ring light the size of a satellite dish. YouTube says ranking still prioritises relevance, engagement and quality, so popularity isn't going anywhere. But "relevant to a precise, constraint-heavy question" is a very different exam from "relevant to two keywords", and the long tail has been revising for it for years without knowing.

The classic funnel has traditionally been spread across about six platforms and one very tired shopper. See it on YouTube, Google it, read retailer reviews, check Amazon prices, ask Rufus a final question, get intercepted by a competitor's sponsored listing, forget what you were buying, buy a phone case.

Ask YouTube, combined with YouTube's shopping plumbing, could squash much of that into a single session. Consideration happens inside one conversation. Purchase is one tap away, thanks to three things YouTube has been building rather quietly:

That third one is what most sellers will underestimate. A review filmed months ago, with an affiliate link politely buried in the description, can become shoppable without the creator lifting a finger. The video simply wakes up one morning with a job in sales. Depending on which YouTube figure you pick, native product tags have driven somewhere between around 50% and over 110% more product clicks than description links alone.

Put it together: conversational search makes old, specific reviews findable. Tagging makes them buyable. Question to comparison to product page, without leaving the building.

On the surface, this looks like pure good news. The transaction still lands on Amazon. Cue the balloons.

It's a bit more complicated.

If consideration moves to YouTube, the job of your Amazon listing changes. A shopper arriving from a comparison table, having watched two reviews and grilled them with follow-up questions, isn't browsing. They've decided. Your product detail page stops being a persuasive shelf and becomes a verification step. Is the price what the video said? Is that variant in stock? Does it look like the thing I just watched someone drop off a table?

Where the video and the listing disagree (new packaging, a price change, a discontinued variant) the friction hits precisely when the shopper was ready to pay. Rather like a restaurant that's brilliant right up until the bill arrives in the wrong currency.

The resurfacing effect also works in both directions, and this is where it gets uncomfortable. Conversational retrieval doesn't distinguish between flattering and unflattering. That brutal review of your previous product version? The comparison where your competitor won on every metric, including vibes? Just as findable. And the shopper has no idea it's two years old. Brands have long relied, consciously or not, on the gentle mercy of old content fading. That mercy appears to have been discontinued (unlike, ironically, the product in the video).

Then there's who benefits. Brands with years of creator coverage are sitting on a back catalogue that can now be put to work. Brands without it have nothing for the system to retrieve. In categories where the answer is assembled from creator reviews, a seller with no YouTube presence isn't ranked lower. It simply isn't in the room.

And then there's Rufus. Amazon has invested heavily in keeping product research inside its own walls. If YouTube becomes the more natural place to ask "which one should I buy?", because the answer comes with actual humans demonstrating the thing rather than a tidy paragraph of summarised text, Amazon may keep the checkout but lose the conversation before it. Along with the data and the ad inventory, which for Amazon is not exactly a rounding error.

Before anyone rewrites their 2027 plan:

It's US-only. For the European half of this newsletter, consider this a postcard from the future.

The shopping features are "experimental". YouTube has shared little on timing or eligibility, which in platform-speak roughly translates to "we'll tell you once we've decided".

Products can go missing. YouTube's docs say Ask YouTube may leave out some tagged products, and it hasn't explained how comparisons are ordered. The taggable Amazon catalogue is also a curated subset, so your product may not make the cut without a request.

Measurement is murky. Creators see aggregate daily clicks and estimated earnings for Amazon tags, with no breakdown by product or video. Linking an Amazon sales spike to a specific resurfaced review will be detective work. If attributing creator marketing already gives you a mild headache, this layer adds a sinus infection.

Ads aren't here. Yet. YouTube says Ask YouTube conversations aren't used to show ads during the experimental phase. "During the experimental phase" is doing an Olympic amount of heavy lifting in that sentence. It would be genuinely out of character for a Google-owned platform to leave a meaningful route to purchase unmonetised forever.

Start with diagnosis, not tactics. US-based teams can type the questions customers actually ask into Ask YouTube, with constraints and use cases rather than product names, and see what comes back. Which reviews surface for your category? Do you appear? Which competitors do? How old is the content?

From there, follow the mechanism:

Check tag eligibility. If your creators are in YouTube's affiliate programme, confirm your key products are in the tag-eligible Amazon catalogue. An untaggable product can't use the shortcut.

Rethink creator briefs. It's a reasonable inference, not a confirmed rule (YouTube hasn't said which signals it relies on), that a meaning-based system has more to work with when a review answers concrete questions out loud: specific use cases, constraints and comparisons.

Audit listing consistency. Make sure your listings match the content most likely to send shoppers to them, especially where products, packaging or pricing have changed since filming.

YouTube's archive has spent years as the internet's largest storage cupboard of product opinions, mostly gathering dust. Ask YouTube is turning it into a working part of the purchase journey.

Which leaves a few questions nobody has answered yet. If YouTube becomes where the decision is made and Amazon where it's completed, who owns the customer, and who gets to charge for access? If a creator's back catalogue becomes a long-lived sales asset rather than a fortnight's worth of content, does paying per post suddenly look mispriced? And if old reviews no longer fade, how should a brand think about the content it made, or didn't make, years ago?

For now, the modest conclusion: it's worth knowing what's in the archive. The algorithm already does.

Speak soon, Jo

P.S. If you've never searched YouTube for your own products the way a customer would, this might be the week. Possibly with a stiff drink nearby.

You can help us!

Spread the word to your colleagues or friends who you think would benefit from our weekly insights 🙂 Simply forward this issue.

In addition, we are open to sponsorships. We have more than 70,000 subscribers with 75% of our readers based in the US. To get our rate card and more info, email us at [email protected]

Jo Lambadjieva is an entrepreneur and AI expert in the e-commerce industry. She is the founder and CEO of Amazing Wave, an agency specializing in AI-driven solutions for e-commerce businesses. With over 13 years of experience in digital marketing, agency work, and e-commerce, Joanna has established herself as a thought leader in integrating AI technologies for business growth.

For Team and Agency AI training book an intro call here.