Brazilian e-commerce is expanding rapidly, establishing itself as one of the most promising digital markets in the world. According to ABComm, revenue is projected to reach R$ 224.7 billion in 2025, outpacing growth in mature markets such as North America and Western Europe. However, Google’s FlashBlack 2025 Edition study, which analyzed the performance of the country’s 31 largest retailers during Black Friday, exposed structural weaknesses: 65% had some kind of technical failure in the shopping journey.
This finding highlights a gap between market potential and the sector’s operational capability, especially in the strategic use of Artificial Intelligence (AI). While consumers already use the technology to guide their purchases and retailers say they plan to invest more in it, practical adoption remains limited.
The following chart makes clear that these failures are not isolated incidents but recurring problems across different stages of the customer journey. This landscape makes the detailed analysis below all the more urgent, showing how AI can be decisive in turning growth into profitability and loyalty.
First Impressions: Weaknesses at the Start of the Shopping Journey
The start of the shopping journey is critical to the success of an online transaction, and in Brazilian e-commerce this moment still suffers from serious failures. During Black Friday, FlashBlack 2025 Edition found that 32% of online stores experienced timeouts — errors that completely block users from accessing the site. This problem, comparable to a physical store with its doors shut during business hours, occurred not only at peak demand but also in other periods analyzed, revealing a structural inability to handle traffic. In addition, 35% of retailers failed to load their main content within the ideal 2.5 seconds, as measured by Largest Contentful Paint (LCP), a metric that tracks when the most relevant page element — usually the hero image or promotional banner — appears on screen. Even though all of them quickly display some visual element to “give the impression” of speed, delays in loading the important content frustrate shoppers and drive up abandonment.
For those who manage to get in and start browsing, another obstacle emerges: visual instability. The study showed that 42% of retailers had high Cumulative Layout Shift (CLS), which occurs when page elements “move around” while the user interacts, leading to accidental clicks or mistaken actions. Combined with the bugs and technical failures reported on 65% of the sites analyzed, this instability turns shopping into a test of patience. Many of these failures stem from “technical debt” — an accumulation of unresolved code and infrastructure issues, usually because the company prioritizes launching new features and campaigns over maintaining its technology foundation. The result is an increasingly fragile platform that erodes customer trust and lowers conversion rates.
Another sensitive point is personalization, which in today’s e-commerce has gone from differentiator to baseline requirement. Only 42% of online stores use browsing history to offer personalized recommendations right on the home page, and just 13% deliver search results tailored to the customer’s profile. In addition, 52% fail to suggest complementary products on the product page, missing one of the simplest and most effective ways to increase average order value. These numbers stand in stark contrast to Brazilian consumer behavior: 66.3% prefer to buy from sites that offer personalized suggestions, and more than half have already used AI tools such as ChatGPT to guide purchase decisions. In other words, shoppers are already used to intelligent interactions, yet most companies still deliver generic experiences.
Ignoring personalization means losing sales not only in the moment but also over the medium and long term. Major players such as Mercado Livre, Amazon and Shopee already use sophisticated recommendation engines that “learn” from user behavior and deliver increasingly relevant interactions. Retailers that fail to keep pace come across as outdated and disconnected from individual needs, pushing customers toward competitors that better understand their preferences. More than a technology issue, this is about staying competitive in a fast-growing market where first impressions — formed in the opening seconds of browsing — can determine the success or failure of the entire shopping journey.
Mid-Journey: The Chasm Between Intent and Discovery
Overcoming the initial performance and stability hurdles does not guarantee the shopper will complete the purchase. In the middle of the journey — the “product discovery” stage — many online stores fail to connect intent with conversion. The biggest problem is site search, which should be the shortcut for high-intent shoppers. According to FlashBlack 2025, 58% of stores do not correct simple typos (e.g., “bluza” instead of “blusa”), 65% do not understand semantic queries — contextual searches such as “outfit for an evening party” — and only 13% offer visual search, a feature popularized by tools like Google Lens. This shortcoming turns what should accelerate the purchase into a barrier, causing ready-to-buy shoppers to abandon the site.
Site search is the “salesperson” of the online store. Just as in a physical store the customer describes what they want and expects to be understood immediately, online they type or speak what they are looking for and expect a fast, accurate answer. Yet most retailers treat this feature as a simple database filter, ignoring the standard set by platforms like Google and AI assistants. Today, consumers search conversationally, using complex descriptions such as “Mother’s Day gift for someone who loves gardening.” The lack of intelligence to interpret these queries reflects not only a technical limitation but also an outdated understanding of modern users, who are already accustomed to smarter, more fluid interactions.
Once the customer finds the product, the next step is the product page — a crucial moment for building trust. Yet only 48% provide complete information, and just 32% invest in immersive media such as videos, 360° views or augmented reality, which help “replace” the in-store experience. Social proof, a decisive factor in purchase decisions, is also underused: although most sites allow reviews, only 10% use AI to summarize them, saving shoppers from reading dozens of comments. Without clear technical details, multimedia content and concise review summaries, an information vacuum forms that increases hesitation and opens the door for customers to buy from more complete competitors.
Finally, the shopping cart, which should be one of the most strategic points of the journey, is still treated passively. 65% of online stores send no notification to remind customers of items left in their cart, despite a 83% abandonment rate in Latin America. Many shoppers use the cart as a wish list or a “save for later” space while waiting for promotions or better terms. This means an abandoned cart is not necessarily a lost sale but a conversion opportunity. By failing to adopt even basic automations to re-engage — something achievable with simple marketing rules and AI — retailers waste highly qualified, low-cost leads and leave revenue on the table.
The Home Stretch: Where Loyalty Is Won or Lost
The customer journey does not end with the click on “buy.” The post-purchase phase — covering logistics and customer service — is when the promise made online must materialize in the physical world. This is where trust is confirmed and loyalty is built. Failures at this stage, however, wipe out the entire marketing and sales investment, turning positive experiences into frustration. In FlashBlack 2025, the data show that the final stage is still riddled with bottlenecks, affecting everything from delivery to customer support. More than just selling, retailers must ensure products arrive quickly, with a transparent process and effective support if something goes wrong.
The first challenge is logistics, especially the “last mile” — the final leg of the product’s trip to the customer. Among the 31 retailers analyzed, 29% do not offer next-day delivery and 71% do not allow customers to schedule a delivery time, something 48% of consumers consider decisive. Only 3 can validate invalid addresses at checkout, preventing errors that cause returns or delays, and 23% fail to provide a clear delivery date or working tracking. In Brazil, where 65% of freight depends on road transport and logistics costs keep rising, delivery has become a competitive battleground. Brands that view logistics merely as a cost miss the chance to turn it into a strategic differentiator that drives loyalty and referrals.
Customer service, especially through chatbots, also shows significant gaps. Of the 26 evaluated, nearly half lack Natural Language Processing (NLP), meaning they cannot understand questions outside a pre-programmed script. Only 2 apply sentiment analysis to detect frustration, another 2 interpret voice messages, and none acts as a shopping assistant or performs simple tasks such as changing a delivery address. Worse, when handing off to human agents, most lose the conversation history, forcing customers to repeat their issue. This contrasts with the expectations of 72% of consumers, who see chatbots as essential, and the 74% who would like them to understand their voice. Retailers that use the technology merely to filter contacts create more friction; those that use it as a proactive tool can turn customer service into an engine of revenue and loyalty.
Finally, digital accessibility remains a blind spot for Brazilian e-commerce. Only 13% of sites achieved an ideal accessibility score, and among apps, just 5 offer features such as screen readers, Brazilian Sign Language (Libras) translation or text-to-speech. This limits the experience for millions of Brazilians with disabilities — a market with significant purchasing potential. Some still see accessibility as a niche or a mere legal obligation, when in practice it expands the total addressable market and strengthens brand reputation. With consumers favoring purpose-driven brands, investing in accessibility is more than “doing the right thing”: it creates competitive advantage, earns loyalty and aligns the business with the expectations of an increasingly inclusive society.
Conclusion: A Strategic Roadmap to Close the AI Gap in 2025
Analyzing the customer journey in Brazilian e-commerce reveals a paradox: a fast-growing market held back by recurring operational failures. Artificial Intelligence (AI) is the most practical and powerful tool for eliminating these bottlenecks, acting directly on performance, experience and loyalty. Today, companies are trying to capture a R$ 224 billion market with what amounts to a “leaky bucket” — losses at the top of the funnel from slowness, instability and lack of personalization; in the middle, from ineffective site search and unconvincing product pages; and at the bottom, from poor logistics and customer service that fails to resolve issues. Although industry leaders plan to increase their AI investments by up to 73%, there is little clarity on how to turn those resources into tangible improvements for the customer.
Closing this gap requires a structured AI maturity roadmap with three levels. The Foundational level fixes the basics, such as typo-tolerant search, abandoned-cart notifications and stability during demand peaks. The Competitive level brings advances such as semantic search, robust personalization and chatbots with Natural Language Processing that retain context. The Market Leadership level pursues innovation, with generative AI shopping assistants, real-time sentiment analysis and multimodal AI to forecast trends and optimize inventory. The table below summarizes the “gap” between what consumers expect and what retailers deliver:
| Area of Opportunity | Consumer Expectations (Based on Surveys) | Retail Reality (Based on the FlashBlack Report) |
| Smart Search | “I want the site to understand what I’m looking for, even if I misspell it or use natural language.” | Only 13% personalize search by profile; 58% fail on typos. |
| AI-Powered Customer Service | 74% would like AI agents to understand their voice; 72% consider AI useful for faster service. | Only 2 of 26 chatbots understand voice; none acts as a shopping assistant. |
| Product Discovery | 52% have used AI (e.g., ChatGPT) to help them shop; 45% have used AI to discover new products | Only 4 of 31 online stores offer visual search; only 2 of 20 use AI to summarize reviews. |
These numbers underscore the urgency of transformation. At every point analyzed, consumer expectations far exceed what retailers actually deliver. Even with mature, accessible technology, the Brazilian market still operates with capabilities that fall short of what it takes to compete globally. Competitors are already moving, consumers have changed their habits and the opportunity is clear: those who rapidly implement this AI maturity roadmap will not only gain market share but also set the standard for the e-commerce experience in the country.