Winning in Amazon's AI Era
July 21, 2026, 1:00PM
35m 0s
Curtis Gerald Ritchie 0:06
Hi, everyone. Thank you for joining the webinar here today. We're just going to
give it maybe 30 seconds to a minute and let a few more people join, and then
we'll get today's session started. How's your afternoon been going, Toby?
Toby White 0:24
Yeah, all good, thanks. A bit colder, isn't it? It's cleared. I've been working
with like fans on it now. My girlfriend was saying that she needs to work with
like a hot water bottle now. It's just kind of changed overnight, isn't it?
Curtis Gerald Ritchie 0:31
Yeah.
Yeah, we're not, we're used to the thirty-twos.
Toby White 0:40
That's quite nice of you there.
Curtis Gerald Ritchie 0:42
No, it's nice being back a little cooler now, but OK, let's get started. So,
good afternoon, everyone, and welcome to our Toucan Talks webinar today. We're
super excited to have you all join us for today's session. If you don't know,
I'm Curtis, our marketing lead here at Toucan.
Toby White 0:48
Yeah.
Curtis Gerald Ritchie 1:02
and we have Toby White, who is Toucan's senior client manager, who will be
talking about how you can win in Amazon's AI era. The session today will run
for around 30 minutes and we'll have a live Q&A at the end of this session.
So if you have a question, please feel free to type them into the Q&A
section at any time during the webinar.
I will collect all of these while Toby is presenting and we will try to get
through as many as possible at the end. So I will now hand over to Toby to take
you through today's session. Thanks.
Toby White 1:34
Brilliant. Thanks, Curtis. Yeah, thanks everyone for joining today. I just want
to say, yeah, big sort of saying a big thank you for joining. As Curtis said,
it's the most sign-ups we've had on a webinar for a while. So I think AI is a
topic on everyone's tongue. So it's brilliant to kind of to have you all here
whilst you're halfway through your day. So yeah, thanks very much. But
When Curtis and the team asked me to kind of present on AI in the Amazon
ecosystem today, there were so many different topics and kind of things that
came to mind, given that AI is constantly evolving and becoming an integrated
and a part in the space. You know, whether it's third party tools being kind of
introduced, which are utilising AI,
or Amazon themselves incorporating AI into the back end or front end of their
systems. AI is kind of changing the landscape fast. But what I wanted to focus
on today is to kind of delve into how conversational AI powered shopping is
rewriting product discovery.
on Amazon and what brands need to do today to succeed in this new era.
So in terms of today's agenda, we're going to discuss the evolution of Amazon
product discovery. So looking at how customers used to find products in Amazon,
how that's changing.
how AI is changing the shopping journey. So specifically the shift from, sorry,
shift to Amazon's AI powered experience, you know, its recommendations and
conversational commerce.
We're going to look at what AI looks for when recommending products. So we'll
go through six core signals that influence whether your products are being
recommended.
And then we're going to look at why traditional Amazon SEO is no longer enough
on its own and the transition to, oh sorry, from a predominantly kind of
keyword focused stance for copywriting to intent, context and answering
consumers' questions across your Amazon listings.
And then finally wrapping up, we're going to go through what brand should do
today. So we'll leave you with five or six practical actions you can take away
today and make your catalogue AI ready and have a competitive advantage moving
forward. So just before I want to start, I just wanted to kind of go through
why
matters. So Amazon has moved AI from a side experiment to the centre of its
shopping experience in barely 18 months. So this isn't kind of a future trend.
It's live and it's at scale and it's already kind of shaping what consumers buy
and how they purchase on the platform.
Some people say that the shift is as big as kind of moving from desktop to
mobile or organic to kind of paid advertising and brands that kind of adapt
early to those shifts win disproportionately. And the same kind of kind of be
said about this AI era where we're kind of in at the moment as well.
Ultimately, the good news is that the fundamentals that win within AI are
clear, kind of accurate, well-structured product information. And then there
are all kind of things which you can kind of control on your listings. So the
game is winnable.
And maybe by the end of the presentation, you'll leave with a clear picture of
how discovery is changing, what AI actually looks at when it recommends a
product, why classic Amazon SEO is still necessary, but it's no longer
sufficient on its own, and a concrete checklist of actions you can start this
quarter to kind of win in this era.
Amazing. So I'm not sure how early everyone's kind of on this call started
adopting and using Amazon, but one thing is clear when we look at the evolution
of Amazon's homepage over the years, and that is that change is constant. So
change is not just happening as well across Amazon's homepage, but it's kind of
across Amazon's front end, back end, you know, their policies and best factors
are constantly changing as well.
and far more as well. So from its humble beginnings to its current sleek design
over the years, Amazon's website layout has adapted and evolved to meet the
needs of its users, showcasing Amazon's commitment to starting a seamless,
providing a seamless shopping experience for its customers. And in 2026, that
means the integration and adoption of AI.
So I don't think these screenshots are particularly clear, but if you were to
zoom in on the 2026 screenshot, you can see Rufus renamed Alexa for shopping
recently in that top left hand corner of the 2026 screenshot. And what that is,
is it's a conversational generative AI powered shopping assistant integrated
into Amazon's app.
and websites to help users research products, read review summaries and compare
items. So what does this change mean for brands trying to kind of win on the
platform then?
So traditionally, for decades, winning on Amazon meant winning the search box,
and that era is potentially ending. So customers are no longer just typing
keywords, they're having conversations, and AI is deciding which five or so
products to put in front of them. And today is about how brands
kind of how about how sorry your brand become one of those five products show.
So to understand Amazon Discovery is kind of the direction it's going. It helps
remember how we used to find things. So the search bar was the front door to
kind of consumers finding your products. They type a few kind of keywords like
dog biscuits and Amazon's algorithms returned a rank
grid of results across multiple pages. Winning in this kind of era meant
rankings, so making sure that your listings have the right keywords
incorporated into the title, the bullet points, and like a strong sales
velocity, competitive price, good reviews, and also obviously like an ad spend
to make sure you're kind of buying the top slots.
Discovery was a funnel with many options, so customers scroll pages of results,
open several tabs, and compared to the listings themselves. Brands' jobs were
to kind of present themselves and be persuasive across a wide net of search
terms. So what did this mean for brands and agencies like Toucan? Well,
Traditionally, some of the ways which we've helped our clients with kind of
this more traditional type of discovery include, but not limited to, content
and SEO audit. So we would often review product portfolios to understand the
kind of the status quo of the listings, score the products with a listing
quality score, and look at
you know, metrics like keyword performance score, which helps kind of gauge the
overall kind of organic reach potential of your products. We'd also obviously
look at copywriting, so writing with a balance of conversion, optimisation and
search visibility across keywords in mind.
hero image optimisation, A/B testing titles and so on, helping clients to
increase their click-through rate, win the click in a busy search engine result
pages. And then PPC, so increasing visibility across the marketing funnel,
different ad types and search engine result page ad positions.
And then finally, making sure that we're kind of monitoring and doing kind of
consistent price analysis and strategic pricing recommendations, including
across Amazon's tentpole events like Prime Day and Black Friday. So whilst all
of those remain important, we need to acknowledge that the ongoing shift in
consumer shopping behaviours
to ensure our optimisations remain optimal on the platform.
And what that means for optimising is understanding the rise of conversational
shopping and the ongoing adoption of Rufus or Alexa as it's kind of being
renamed.
So firstly, how is this rise in kind of the use of AI on the platform and its
integration to the platform come about? So shoppers are bringing their AI
habits from the rest of their lives. So I'm sure everyone on the call has
probably used an AI system like ChatGPT or Gemini or
My personal favourite is Claude. But we shoppers are bringing that kind of
those AI habits from their personal use or their use of AI at work into how
they research purchases on Amazon. So they increasingly expect to describe a
need, not just to kind of guess the right keyword to show their desired
products. So
Ultimately, research is moving upstream. People are spending longer within AI
tools exploring, you know, help me figure out what I need and arriving at the
point of purchase already kind of informed rather than a prolonged
consideration phase on Amazon where you're scrolling different pages and kind
of opening different tabs.
with different when comparing different products. And ultimately, trust is
being delegated. So a meaningful share of consumers are comfortable letting AI
summarise reviews and do the kind of comparison phase of different listings for
themselves.
Just to kind of throw a few stats out there on this kind of shift in the use of
AI across e-commerce platforms, including Amazon. So in a survey of over 4,000
consumers, 39% said that they had already used AI to shop a behaviour which
barely existed a year or two ago.
and 79% of those in the survey felt more confident in what they bought, having
used AI to find that product. More than 300 million customers have used
Amazon's AI assistant during 2025, and analysts expect that roughly 1/4 of
shoppers will use AI chat.
while shopping in 2026. So about 1/3 of US consumers say that they'd let AI
actually going to make that purchase for them. And in the US, they are actually
testing that functionality out with Amazon Rufus. So as an auto buy feature,
integrated into Rufus or you want to call it Alexa. So they're testing that out
in the US. And so with a lot of kind of the different things we see on Amazon,
if it's being tested in the US, I'm sure at some point it will also be tested
and maybe integrated into the UK as well.
So just to reiterate the kind of the shift we're talking about for you all
then, so instead of typing keywords and scrolling, customers are increasingly
asking kind of questions in natural language, such as kind of what I've put in
this screenshot here, so what to buy for a home spa day, and they expect a
direct answer.
Amazon's AI assistant sits inside the app or website, and it answers kind of
questions, compares products, and it recommends a short list. In terms of that
short list of recommendations, from our observations, this can this is like
often around 5 products. So, rather than you know scrolling through 50 or so
results, you see
in your search result pages, you know, the kind of the ultimately the
recommendations kind of throw up five rather than the 50, so it's kind of it's
a much smaller net being cast.
And then in fact, the search engine results pages can be skipped altogether. So
if you were to start, you know, asking Rufus some questions on the Amazon
homepage, you could kind of click through on one of those recommendations and
it will take you straight through to the product detail page. So search
the kind of search engine results pages on, or the search result pages on
Amazon can kind of completely be skipped altogether. So ultimately, this
collapses the funnel. AI does the scrolling, comparing on the customer's
behalf. The shopper sees a curated handful of products with reasons attached.
for why each of those is chosen. So in the screenshot here, you can see some of
the short reasonings for the products underneath the listings. So for example,
the Sanctuary Spa kit was showcased because it has like a mixture of body wash,
scrub, lotion,
and asleepness. So all sounds kind of lovely for an at-home spa day.
Um...
And ultimately, so what does this mean for the brand? So the battle from
shifting is shifting from appearing somewhere on the search results page and
being kind of one of 50 on that search results page to being one of five or a
much smaller kind of collection, which the AI is recommending. So
In short, the shelf just got kind of dramatically smaller.
So AI is not just showing on the homepage. It doesn't stop there. It's also
being shown across kind of all other pages on Amazon, including the product
page itself.
So in these screenshots, you can see that Rufus is offering consumers the
ability to look at the price history. So there's a look-back window of up to
one year for comparing kind of like the peaks and troughs of the historical
kind of pricing for that product. And you can also answer, it also answers kind
of live questions for you, such as,
Is this treat good for dogs with sensitive stomachs? Or, you know, are they
made in the UK? So it's pulling those answers from the listing elements and
reviews. And if the answers aren't there, sorry, if the answers are there, then
you, you know, you win in the moment and the consumers' questions are answered.
But if your listings don't have that ability to answer the questions which they
are asking,
you risk losing the consumer to a competitive product.
So in short, they've woven generative and agentic AI through the whole journey
on Amazon from discovery, comparison, decision, and increasingly the purchase
itself. Like I said, it's being tested out in the US with like auto buy and the
ability to actually kind of allow AI to make the purchases for you.
First, the AI assistant answers questions, so it compares options and
recommends products with reasons attached. So it's not just kind of like a
static search box is what we were kind of used to previously, but it's an
active kind of shopping advisor working on your behalf.
Secondly, we have AI recommendations, so it pulls out review highlights, tells
the shopper why this product, and it kind of builds comparisons for consumers
on the fly.
Third is it can provide kind of personalised journeys. So two shoppers can ask
the same kind of question and it might get slightly different answers or
products kind of shown. And that's based on their kind of search history and
purchase history on the platform. And so the shelf can be slightly different
for everyone.
And then fourth is the conversational commerce. So this is a big one. The AI
can build, track price history like we showed you in the screenshot previously.
And like I said, it can, they're testing all the auto buy feature.
in the US. So when a product hits a certain target price, which you can kind of
discuss with the AI, it will do that automatic purchase on your behalf.
So if an AI is choosing roughly 5 products, so like I said, in that kind of the
recommendations, once you've asked questions, we typically see that it does
like a minimum of five products, sometimes slightly more. But if AI is choosing
roughly 5 products it recommends, the obvious question is, what is it looking
at?
And the answer is that is ultimately it's looking for evidence. So there are
kind of six different kinds of evidence that we perceive matters most. And
those are product content quality. So the AI reads your title, your bullet
points, description, you know, content like your A plus content, the question
and answers on your PDPs.
And they use that as source material to answer the questions. Complete,
specific, factual content gives it more to work with whilst, you know, if your
products had like thin or vague content, it can kind of leave the AI guessing.
It ultimately, you know, rewards clarity of use case. So if you're
If your product was a backpack and you were explicit about who is it suitable
for, so it might be best for commuters and frequent flyers, but not ideal for,
you know, people who have laptops over a certain size. You know, if you can be
specific, that can help the AI match your product to the right kind of
consumers.
and the right consumers asking those questions.
Secondly, it will kind of pull on the product attributes and metadata. So this
is kind of structured back end data, which is often like hidden from the
product pages itself. So the attribute fields, specifications, variation
relationships, and these are now all primary levers.
to help kind of your, the AI find your products, promote your products. So
having missing or inconsistent attributes can make your products invisible to
filtered intent-based queries, whereas populating, you know, relevant
attributes with information
such as, if it's obviously applicable to your product, you know, being
dishwasher safe or suitable for sensitive skin, these kind of types of
attributes which Amazon allows you to populate, then the AI can kind of help
push you towards queries which are asking
which those attributes aren't stuff. Whereas if you don't, again, you'll
struggle to be shown in the recommendations.
Then customer reviews and sentiment. So reviews are now evidence, the AI mines
for themes, sentiment, and specific use cases, and it cross references your
claims against them. So if you claim a product is leak-proof and the reviews
say otherwise, the AI will kind of surface that gap.
So review volume, recency, rating and the substance of reviews all feeds into
how confidently the AI will recommend to you.
Then moving on to product availability and pricing. So the assistant factors in
price competitiveness and price history. And like I was saying before, it can
kind of show up to a year of price history. And then also availability when it
comes to what to recommend and when to kind of
nudge of product towards a purchase. So out of stock or erratically kind of
priced products are poor recommendation. I was have poor recommendations with
AI was less likely to surface them.
Brand authority and trust signals, so consistent brand information, credible
third party signals and coherent presence across all the web will help your
deal with the with AI trust. And therefore kind of we recommend your product
pools, or sorry,
the AI will look and pull from beyond Amazon as well. So you want to make sure
that, you know, all your touch points online kind of have the right information
or are kind of populated to the best of your ability. And trust signals reduce
the AI's risk of recommending something that disappoints, which is exactly what
it's optimising against. So
Just worth kind of keeping in mind as well. Lastly, then, so shopping history
and personalisation. So the assistant blends the individual shoppers past
purchasing, purchases, browsing and stated preferences into what it recommends.
So two customers asking it as a go.
and identical questions can get slightly different results. And so something to
keep in mind with. But it also gives you an opportunity in terms of loyalty. So
if someone has bought from you before, they're more likely to be surfaced to
them again when they're kind of searching and asking queries on using AI.
So yeah, repeat purchase and subscribe and save behaviour compound into future
visibility. So some to kind of, yeah, all of those things to kind of keep in
mind when you're thinking about the kind of impact and where you can kind of
optimise your listings and put yourself in the best place for
AI recommendations.
So one thing I just want to be clear on is that traditional Amazon SEO isn't
dead and everything that you focused on previously still matters, but on its
own is now kind of the price of entry and not the path of necessarily winning.
Your optimisation strategies need to shift that focus from solely kind of
keyword optimisation
to also factoring in intent optimisation. So whilst classic SEO optimises for
words and customer types, intent optimisation asks what is the customer
actually trying to achieve, and does my product clearly answer that? The AI
interprets meaning, not
just keyword phrase matches, it maps a messy human request to kind of the
products that genuinely fit that kind of that criteria.
So my suggestion would be to keep your kind of keywords, but layer an intent on
top of that. So cover the job situation outcomes your products delivers in
plain language as well. So yeah, stop kind of just being keyword focused. And
the question we need to ultimately ask, you know, previously was, am I ranking
for these keywords? Whereas the question which we need to ask ourselves now,
is if a consumer described their need out loud, could an AI confidently
recommend your product and back it up with the kind of the content and reviews
and all those different elements that we spoke about previously?
Amazing. So, ultimately, I think this is probably the most important slide, and
it's what brands should do today to not fall further behind in the AI era. So,
firstly, it's important to say that none of this is, you know, rocket science,
it just requires you to kind of do the fundamentals deliberately.
with AI as your reader in mind. So the five moves that you can kind of start
this quarter to make some real progress and build some traction in this area
would be, sorry, six moves, not five. So the first one would be to audit your
PDP content. So
Review every product detail page against one test, and that is, could the AI
answer a shopper's real question using only this page, and if it can't, you
know, you wanna fill the gap.
So whether it's fit, use cases, compatibility, what's in the box, or who's it
not for? You know, think of all the like, bring some all of these different
questions that consumers might ask when looking at your product page. And then
I would suggest prioritising, you know, top to bottom, like your highest
revenue.
generating or highest traffic ASINs, you know, ask those questions on those
PDPs and ultimately try and answer those questions in the content which we've
spoken about.
Next one would be to improve your product attributes and metadata. So complete
every relevant backend attribute and specification accurately.
treat any empty attributes you find as kind of lost recommendations, and then
check variation relationships and structured data for consistency. So for that
one, kind of the quickest thing you can do is to download on Seller Central,
you could download your category listings report, and then on vendor, you can
kind of just go onto your catalogue and
do bulk exports and then you'll be able to see kind of like the status quo in
terms of, you know, what's populated in the attributes and where are the gaps
and what can we kind of populate those gaps with.
The next thing you could do is obviously strengthen your review strategies. So
try and drive more recent substantive reviews through compliant means,
obviously. So obviously one of those ones which we often recommend to our
clients is the Vine programme. And then you've also got kind of options in
terms of post-purchase follow-ups.
and ultimately, you know, trying to sell great products so that you get good
reviews in the first place. But recency and relevance now feed into visibility
with the AI, so it's not just conversion. So yeah, thing to keep in mind as
well. Ultimately, you want to read your reviews for themes that the AI will
extract.
and make sure your content honestly reflects those. And then you want to build
more informative content. So create content that answers questions and explains
the content. So making sure you're going to be using rich A+ content modules,
having enough thorough Q&As and comparison.
friendly kind of detail onto the PDPs if you can. So obviously you want to try
and inform first and then sell second almost.
anticipate the comparisons that AI will make with kind of other products in
your market as well, and give it the facts to put you in a kind of a good light
and make your products ones that it recommends.
And...
And then create consistency across channels. So make sure that your brand and
product information is accurate and aligned wherever they might read it. So the
AI on Amazon does have the ability to, you know, look beyond Amazon listings
itself. So making sure that your Amazon listings, your website, other retailer
pages or third party sources are kind of consistent.
And yeah, if there were inconsistent specs or claims across the web, then that
can kind of create doubt. And so consistency builds the trust and that earns
the recommendations. And so yeah, a quick win there would be to kind of do a
spot check across your top products.
compare Amazon to your own sites and any kind of other third party sites or
marketplaces and just see if there's any mismatches. But the ultimate goal
there would be to have consistency.
Amazing. So to wrap up, some of the key takeaways from today's session. So
discovery is ultimately changing. Customers are moving from typing keywords to
asking questions. And AI is curating a short list of answers, often about 5
products rather than the multiple pages of product.
that we've seen in typical kind of search result pages.
AI recommends on evidence, so content quality, structured attributes, reviews
and sentiment, availability and pricing, all the things that we discussed are
the main influencing factors.
Keywords aren't enough, so optimise for intent, questions, and context as well.
right so that an AI could confidently recommend you from your page and reviews
alone. Accuracy beats hype, so the AI cross-checks your claims against reviews
and the wider web, so you know, be honest, be complete and consistent. That is
the ultimate kind of winning factors.
And ultimately start now with the basics. So audit your PDPs, complete your
attributes, strengthen your reviews, add informative content wherever you can,
and try and kind of stay consistent across channels. And finally, you know, the
brands that win in this new era with
that AI's new era won't be the loudest or the cleverest kind of gaming search.
They'll be the clearest. So the ones whose products are well described, well
reviewed, that Amazon has every reason to recommend will ultimately win. And
all of that is entirely within your control. So the best time to start is now.
And hopefully off the back of this deck, you're going to have a bit more
clarity as where to start and what you can do to kind of influence your
products showing in the recommendations.
Yeah, that's everything.
Curtis Gerald Ritchie 30:38
Thanks.
Thanks so much for that, Toby. With all the talk and the new AI features coming
out near every month, it's really important to keep up to date with everything.
So that was a super insightful session. We'll quickly jump in to the Q&A
section. We'll get a few questions before we wrap up today's webinar.
Toby White 30:42
That's right.
Curtis Gerald Ritchie 30:59
So the first question we have here would be how would a brand actually test
whether the AI is recommending them today?
Toby White 31:08
Yeah, so it's a good question. There's not currently a report which kind of
attributes the, you know, the questions which are resulting in like a product
sale or anything like that. Maybe that's something which Amazon will look to
produce in the future. But for now, you know, I think the main things you can
do is just simply kind of
logging on, maybe kind of go onto a private browser so it doesn't have that
historical kind of data. And, you know, think of the industry you're in, think
of the products you sell and what kind of questions consumers would ask. So
yeah, if it's a pet brand, you know, what kind of, think of that questions on
breed types or
you know, making sure different allergens are flagged in the data and things
like that. But yeah, unfortunately, there's no specific report for it at the
moment. So you kind of just have to go in, ask kind of potential questions
which you think are relevant and see if your products kind of your show.
Um, so yeah.
Curtis Gerald Ritchie 32:10
Thank you for that. We've got two questions here from Adam. One was on tracking
kind of any impact changes, but you kind of just answered that. There's not
really too many ways just at the moment, but hopefully Amazon will come out
with something potentially soon. But one of his other questions is if you were
to focus on one action above all others, what would it be?
Toby White 32:33
Yeah, that's a good question again. So I think the one action which gets kind
of forgotten about is probably the product attributes one. I think because it's
a lot of the attributes are in the back end, so it's not kind of the glamorous
kind of pretty content you see on the PDPs, it can often get kind of overlooked
and forgotten about.
Um...
you know, if you download the sheets, you know, the spreadsheets are relevant
to, and the attributes are relevant to each of the categories which your
products are in. So, you know, there's going to be lots of different attributes
specific to all the different products which the AI can draw upon. So much I
mentioned.
you know, like breed type and such, but yeah, I think that's the one which
probably gets forgotten most and one which is, yeah, undervalued.
Curtis Gerald Ritchie 33:19
Yeah.
Thanks for that. And we'll finish up with one more question. What would be one
of the most common mistakes you're seeing on listings right now?
Toby White 33:39
Well, yeah, the famous one previously is probably being like keyword stuffing.
I think, yeah, people kind of getting more conscious on the avoiding that. But
historically, it's being keyword stuffed. But now it's probably the mistake is
to not find that right balance between, you know, having the relevant keywords
and also optimising for AI. So yeah, previously probably keyword stuffing and
now I would say it's yeah, not having that good balance between all the
different kind of areas which we need to focus on optimising on at the moment.
Curtis Gerald Ritchie 34:05
Mhm.
Thanks for that, Toby. Okay, I'm aware we've run slightly over, folks, but we
will wrap up there. So I just want to say a massive thank you for everyone
joining our Toucan Talks webinar today, and we hope today's session has been
super useful for you. If you would like to get in touch with Toby regarding
today's session, or you would like to see how Toucan
Toby White 34:20
So.
Curtis Gerald Ritchie 34:35
can help with winning in this Amazon AI era. Toby's email is on the screen
there, toby.white@toucanecommerce.com. So thanks again and we hope to see you
at another Toucan Talks event, whether that is online or in person in the
future. So I hope everyone has a great rest of their week and thank you all for
coming.
Toby White 34:55
Thanks, everyone.
stopped transcription
