- Should you invest in AI discovery or in reviews? What's the right priority?
- What does the data say: how do people really behave after an AI recommendation?
- How widespread is the use of AI in purchasing in Italy?
- Three concrete scenarios: what changes for an e-commerce, a professional firm, and a manufacturing company?
- Which touchpoints should I optimize to convert traffic coming from AI?
- How do discovery channels compare: AI, traditional search and other tools?
- What are the risks if I focus only on AI visibility and neglect conversion touchpoints?
- Why it's being discussed now: is the buying journey really changing?
- How to get started: the steps to rebalance your strategy
- How do I know if the strategy is really working?
- So, is it worth changing strategy for your company?
- How we see it at SHM
- Frequently Asked Questions
- AI (artificial intelligence) wins at product discovery, but reviews and purchase information close the sale: only 5% buy right after an AI recommendation (source: Feedvisor 2026).
- 33% of shoppers check reviews after an AI suggestion and 22% compare prices: optimizing these touchpoints is worth more than investing in AI visibility alone (source: PartnerCentric 2026).
- In Italy, 40% of online adults already use AI tools and 53% of marketplace buyers read reviews before purchasing: the multi-touchpoint journey is already a reality (source: Netcomm 2026).
- The practical priority for SMEs is to rebalance the budget: first take care of reviews, product pages and return policies, then invest in AI discovery strategies.
Only 5% of shoppers buy right after an AI recommendation. The other 95% go check reviews, prices and return policies before opening their wallets. If your product page doesn't hold up to that scrutiny, AI just gave someone else free advertising.
Should you invest in AI discovery or in reviews? What's the right priority?
It's not an either-or choice: AI brings qualified traffic, but reviews and purchase information decide whether that traffic converts. For most SMEs, the priority is to consolidate conversion touchpoints first, then amplify AI visibility.
Imagine a customer asking ChatGPT which supplement to buy. The AI suggests your brand. They open Amazon, read the reviews, see that the latest one is eight months old and that someone is complaining about shipping. They buy from your competitor. AI found you, but your product page lost you the sale.
This is the scenario that emerges from the research published by Digiday and from Feedvisor and PartnerCentric data. Before shifting budget, it's worth understanding where you stand right now.
| Your situation | Priority | why |
|---|---|---|
| Few reviews, or outdated ones | No, for now | AI mentions you, but customers don't convert: fix this first |
| Incomplete or inaccurate product pages | No, for now | AI reads the same sources your customers do: wrong information penalizes you twice |
| Unclear or missing return policies | No, for now | 33% of shoppers check them before buying |
| Up-to-date reviews, complete product pages, clear policies | Yes | Now it makes sense to invest in AI visibility and GEO (optimization for AI engines) |
| New brand with no review history | mid | Build social proof first, then scale discovery |
What does the data say: how do people really behave after an AI recommendation?
After an AI suggestion, 33% of shoppers go read reviews, 22% compare prices and 19% search for the product on Amazon or another marketplace. Only 15% visit the brand's website directly.
PartnerCentric's numbers (2026 survey) show a clear buying journey: AI opens the door, but customers still go the full round before buying. Even those who say they trust AI a lot don't skip verification: 42% of these shoppers still want to read reviews before acting.
One figure stands out: only 5% buy immediately after an AI recommendation, according to Feedvisor. 86% of American online shoppers who used AI to search for products verified the suggestion on another source before buying, according to the Trust in AI Commerce Report by Product.ai (survey of 1,463 shoppers, April 2026).
The point isn't that AI doesn't work. It's that it works at a different stage than the one where many companies are investing.
How widespread is the use of AI in purchasing in Italy?
40% of Italian online adults already use AI tools regularly, and over 13% use them in the pre-purchase phase. The Italian market is following the same trajectory as the American one, a few months behind.
According to the Netcomm AI & Agentic Commerce 2026 White Paper (Comscore/Sensemakers data), the share of Italian online adults who use AI regularly doubled in one year: from 20% in 2025 to 40% in 2026. Among 18-to-34-year-olds, it reaches 49%.
Among the main uses in the pre-purchase phase, according to Netcomm NetRetail 2026: comparing prices and products (67%), getting personalized suggestions (66.8%), summarizing reviews and technical information (61.2%).
Here's a surprising number: according to the Netcomm Observatory with Circana 2026, ChatGPT, Gemini and Copilot influenced the last purchase of 32.7% of Italian consumers. That's close to influencers (33.7%) and higher than TV advertising (27.7%). AI is no longer a niche phenomenon: it's already a mass-market touchpoint.
Three concrete scenarios: what changes for an e-commerce, a professional firm, and a manufacturing company?
The mechanism is the same in all three cases: AI brings visibility, but conversion depends on the quality of the information available online. Only the type of "review" the customer looks for changes.
Illustrative examples: numbers are meant to show the reasoning, they are not real customer data.
| Company type | How AI enters the journey | What the customer checks next | What to optimize first |
|---|---|---|---|
| Cosmetics e-commerce | ChatGPT recommends the product for a specific skin concern | Reviews on Amazon and on the site, ingredients, return policy | Detailed product pages, post-purchase review collection, visible free returns |
| Tax consulting firm | AI answers "who can help me with my VAT return in Milan" | Google My Business, Google reviews, a site with case studies | Up-to-date Google profile, client testimonials, clear service pages |
| B2B components manufacturer | AI cites the brand in an answer about technical specs | Spec sheet on the site, certifications, delivery times | Downloadable technical documentation, product FAQs, a quick contact page |
For e-commerce, the problem is often the gap between the moment AI cites the brand and the moment the customer lands on the product page. If that page has three old reviews and no real photos, the customer bounces. For a professional firm, verification almost always goes through Google Maps and the star rating. For a B2B manufacturer, the customer looks for the PDF spec sheet: if they can't find it, they call your competitor.
Which touchpoints should I optimize to convert traffic coming from AI?
Reviews, accurate product pages, clear shipping and return policies: these are the four elements customers check after an AI recommendation, in the order they check them.
According to the Netcomm Focus Digital Marketing & Channel Strategy 2026, a shopper consults 4 touchpoints on average before buying: AI and chatbots, offline retail media, the brand's site and reviews. Reviews are consulted by 53.1% of those who buy on Italian marketplaces.
The idealo Italian E-commerce Report 2026 adds that detailed product information influences 43% of users and reviews affect 39% in the decision-making process. These aren't marginal figures: they're the two most cited factors after price.
- 1 · Step 1Review auditCount how many reviews you have, when the latest one was posted, and your average rating on every platform where you sell. Under 4.2 stars or fewer than 20 recent reviews: top priority.
- 2 · Step 2Update your product pagesEvery page should answer the questions a customer would ask an AI: ingredients, materials, compatibility, dimensions, use cases. AI reads the same pages.
- 3 · Step 3Make shipping and returns visible33% of shoppers say shipping and return policies often change what they buy. Highlight them on the product page, not just in the footer.
- 4 · Step 4Set up systematic review collectionPost-purchase email, QR code on the packaging, SMS request: choose the right channel for your customer and automate the process.
- 5 · Step 5Monitor AI mentionsSearch for your brand on ChatGPT, Gemini and Perplexity using the questions your customer would ask. Check that the information is accurate and up to date.
How do discovery channels compare: AI, traditional search and other tools?
AI is already the most useful discovery tool for 35% of American consumers, versus 14% for traditional search engines. But the traffic it generates still has to be converted with the same tools as ever.
According to Spangle Insights (July 2026), AI referral traffic to retail sites grew 62% year over year. AI has become the preferred starting point for product research, but it hasn't replaced conversion touchpoints yet.
| Channel | Journey stage | Strength | Current limitation |
|---|---|---|---|
| AI (ChatGPT, Gemini, Perplexity) | Discovery / initial research | Contextual answers, quick summaries | Doesn't close the sale: only 5% buy right away |
| Google Search | Search with specific intent | High intent, direct click | Share of discovery declining versus AI |
| Marketplaces (Amazon, etc.) | Comparison and purchase | Reviews, prices, shipping all in one place | Thin margins, platform dependency |
| Brand website | Verification and direct purchase | Full control over information and margin | Only 14.8% get there after an AI recommendation |
| Creators and influencers | Social proof and inspiration | High perceived trust | Hard to scale and measure |
What are the risks if I focus only on AI visibility and neglect conversion touchpoints?
The main risk is funding your competitor's discovery: AI cites you, the customer finds you, but they buy from whoever has better reviews and clearer information.
AI reads the same sources your customer does: reviews, product pages, retailer pages, creator content, prices, return policies. If those sources are inconsistent or incomplete, AI doesn't fix them: it amplifies them. As Stephanie Harris of PartnerCentric put it, "when signals conflict, AI doesn't smooth them over: it surfaces the problem sooner."
A second risk concerns budget. Investing in GEO (generative engine optimization, the strategy for getting cited by AI) without first strengthening your conversion touchpoints is like advertising a store with broken windows. Traffic arrives, but it doesn't come in.
If you're already running paid campaigns on AI channels, as described in the article on ChatGPT Ads and the budget testing phase, you know measurement is still complex. Adding a weak conversion touchpoint makes everything harder to read.
| Error | Consequence | How to avoid it |
|---|---|---|
| Investing in GEO without taking care of reviews | AI cites you, the customer doesn't convert | Review audit before any AI investment |
| Generic or inaccurate product pages | AI gives wrong information about your brand | Update every page with precise, current technical data |
| Hidden or missing return policies | The customer drops off at the decisive moment | Put shipping and returns front and center on the product page |
| Tracking only AI traffic, not conversions | You don't know if AI is bringing real value | Track the full journey: from where the customer comes from to where they buy |
| Ignoring marketplaces in your AI strategy | 19% of customers go to Amazon after an AI recommendation | Optimize your marketplace listings too, not just your site |
Why is everyone talking about it now: is the purchase journey really changing?
Yes. Referral traffic from AI to retail sites grew 62% in a year. AI is already the most-used discovery touchpoint for a growing share of shoppers, and the Italian market is accelerating.
The shift isn't gradual: it's a break. Until 2024, Google was the near-universal starting point for product search. In 2026, 16% of shoppers start with ChatGPT, according to Feedvisor. And Feedvisor's 2026 E-commerce Consumer Behavior Report (a survey of 1,006 US consumers, August 2026) shows that 55% used AI to research or compare products in the past six months.
If you run paid campaigns, you already know the landscape is fragmenting. Our article on priorities and budget allocation by channel in Google Ads shows how this fragmentation is already changing investment choices. Adding AI as a discovery channel means rethinking the entire customer journey architecture.
The critical issue is seasonality. With year-end sales peaking, companies that enter the season with up-to-date reviews, accurate product pages and clear return policies will be easier for AI to recommend and easier for customers to choose. Those who show up unprepared hand visibility to their competitors.
It's also worth looking at what's happening with the drop in searches during the Black Friday period: traffic is being redistributed across new channels, and AI is one of them.
How to get started: the steps to rebalance your strategy
Start by auditing what you already have, not with new investments. Budget only shifts once your conversion touchpoints are solid.
- 1 · Week 1Audit reviews and product pagesCheck ratings and review counts on every platform. Identify product pages with missing or inaccurate information. Prioritize by traffic volume.
- 2 · Week 2Optimize critical touchpointsUpdate your priority product pages. Make shipping and return policies visible. Set up an automated post-purchase review collection flow.
- 3 · Week 3Check your AI presenceSearch for your brand on ChatGPT, Gemini and Perplexity using your typical customer's questions. Note what is said and whether it's accurate.
- 4 · Week 4Fix incorrect informationUpdate your website pages, marketplace listings and press releases that AI might read. Authoritative, up-to-date sources improve the quality of AI answers.
- 5 · Month 2+Scale AI discoveryOnly now, with solid conversion touchpoints, does it make sense to invest in GEO and campaigns on AI channels. Monitor the full journey, not just traffic.
How do I know if the strategy is really working?
Looking at the traffic that comes from AI isn't enough: you need to follow the customer all the way to purchase and measure the quality of the touchpoints in between.
| Indicator | What it measures | Positive signal |
|---|---|---|
| Average review rating | Perceived quality of the product/service | Above 4.2 stars on all major platforms |
| Number of recent reviews (last 90 days) | Freshness of social proof | At least 5-10 new reviews per month for each key product |
| Conversion rate from AI traffic | Effectiveness of conversion touchpoints | On par with or higher than the rate from other organic channels |
| Bounce rate on product pages | Quality of the information available | Dropping after product page optimization |
| Accurate mentions in AI engines | Quality of the information AI provides about the brand | Correct, up-to-date information on ChatGPT, Gemini, Perplexity |
| % of purchases verified on another source | Reliance on secondary touchpoints | Trackable via post-purchase surveys or heatmaps |
A common mistake is measuring only inbound traffic from AI without tracking what happens next. If the customer arrives from an AI recommendation and then goes to Amazon to buy, your site lost the sale even if traffic grew. You need to track the full journey, not just the first click.
If you run paid campaigns, this ties into the reasoning already described in our article on ChatGPT Ads click-to-chat and conversion without site traffic: new AI formats change how the value of an interaction is measured.
So, is it worth changing strategy for your company?
Worth it if you already have solid reviews, up-to-date product pages and clear return policies: investing in GEO and AI visibility can amplify an already strong foundation and bring in extra qualified traffic.
A middle path you have some reviews but your product pages are incomplete or your purchase policies aren't very visible: work in parallel, optimizing conversion touchpoints while you start monitoring your AI presence.
Not yet if you have few reviews, generic product pages or no return policies: every euro spent on AI discovery now risks sending traffic that converts elsewhere. Consolidate first, then scale.
How we see it at SHM
For us, AI is the means and the customer is the end. This data confirms it almost like a textbook case: AI brings the customer to the doorstep, but it's the quality of the information available — reviews, product pages, purchase policies — that decides whether they come in or turn on their heels.
We believe that simplifying digital is the highest form of expertise. And here, simplifying means not chasing the newest technology before consolidating what already works. The companies that convert AI traffic best in 2026 won't be those that invested the most in GEO, but those that built a consistent, reliable ecosystem of information across every touchpoint.
At SHM Studio we work on this with our clients: from the strategic integration of AI into marketing processes, to optimization for AI engines (GEO), to building websites that hold up to the customer's scrutiny after an AI recommendation. If you want to figure out where to start for your company, get in touch.
You can explore all our articles on this topic in the Marketing and ADV performance section of our blog.
Frequently Asked Questions
Not entirely yet, but the share is significant: 16% of shoppers start their product search on ChatGPT, according to Feedvisor 2026. AI is becoming the preferred starting point for a growing slice of consumers, especially for complex or comparative searches. Google remains dominant, but its share of discovery is down from last year.
It depends on how solid your conversion touchpoints are. If you have up-to-date reviews, complete product pages and clear return policies, it makes sense to start working on AI visibility. If those elements are missing, investing in GEO now risks bringing in traffic that doesn't convert. The right sequence is: consolidate first, then scale.
The simplest way is to search directly on ChatGPT, Gemini and Perplexity using the questions your typical customer would ask. Search for your brand name, your main products and the categories you operate in. Check that the information is accurate and up to date: if AI reports wrong data, update the website pages and marketplace listings it uses as a source.
Yes. AI engines read the same sources your customers do: product pages, marketplace reviews, comparison sites, editorial content. Numerous, recent reviews on Amazon improve both direct conversion on the marketplace and the quality of the information AI provides about your brand.
For reviews, you'll see results in 4-8 weeks if you set up a systematic post-purchase collection flow. For AI visibility, it takes longer: AI engines update their knowledge at varying frequencies, and some sources get indexed faster than others. A realistic plan allows 2-3 months to see a measurable improvement.
Yes, with a few differences. In B2B, the "review" a customer looks for after an AI recommendation is often a data sheet, a certification or a case study. The mechanism is the same: AI opens the door, but the quality of the information available online decides whether the customer moves forward. In B2B too, auditing your online sources is the first step.
AI puts you in the shop window, but your reviews decide whether the customer comes in or walks past.
Related articles
Discover more articles exploring similar topics, selected to offer you a more complete and stimulating perspective. Each piece of content is carefully chosen to enrich your experience.
