Trustworthy Shopping Experience Report

Global retail has evolved—connecting billions of customers with millions of selling partners across dozens of borders at unprecedented speed. Customers can shop through a vast selection of products and goods from around the world, delivered right to their doors faster than ever. Businesses can easily reach regions and customers they never could have before, achieving growth in years that previously would have taken decades.

This creates extraordinary opportunities and equally sophisticated challenges. At Amazon, we’ve built the systems, collaborations, and expertise to meet these challenges head-on. These systems, while advanced, are never done improving. They are in a constant state of change, building upon decades of learnings, billions of data points, and a constant flow of feedback to maintain the most customer-centric shopping experience in the world.

Four pillars guide this work:

• Robust proactive controls protect customers and legitimate selling partners through comprehensive verification systems and compliance programs. From seller identity verification to product safety testing and proactive intellectual property protections, this multilayered approach stops issues before they reach customers.
• Innovative tools and technology combine advanced technology with human expertise to stay ahead of evolving threats. Multimodal systems analyze billions of signals simultaneously—from visual elements and textual content to seller behavior and supply chain patterns—revealing connections that isolated analysis would miss. This comprehensive view allows us to anticipate risks before they materialize or impact customers.
• Holding bad actors accountable through dedicated enforcement teams and global collaborations across the public and private sector. These teams work alongside law enforcement, brands, and industry leaders to combat organized retail crime, counterfeit operations, refund fraud, and scam networks—dismantling criminal infrastructures at their source.
• Protecting and educating consumers delivers crucial information about anti-counterfeiting, scams, and product safety at the moments people need it most. Collaborations with industry leaders and government agencies worldwide ensure consumers have access to knowledge that protects them—whether that’s safety guidance after disasters, scam prevention during peak shopping periods, resources helping parents navigate overwhelming choices, or anti-counterfeiting campaigns educating the next generation about the importance of purchasing authentic products.

The challenges facing global retail are real. But so is the infrastructure we’ve built to address them.

Determining who can sell in our store

Many retailers focus primarily on verifying seller identity at registration. We’ve long done that too—but we also know bad actors are sophisticated and can find ways to circumvent one-time checks. So, for many years, we’ve used continuous monitoring systems that track signals throughout a seller’s journey to identify and respond to potential risks as they emerge.

Our systems verify new sellers through a combination of advanced technology and expert human review both upfront at registration and reverifying, if necessary, throughout their tenure. Our goal is to make it straightforward for genuine businesses to set up a selling account, but for bad actors attempting to misrepresent themselves, they encounter intentional friction at every step. When sellers apply to sell in Amazon’s store, they must provide government-issued IDs, business credentials, bank statements, and proof of address—cross-checked against internal and external databases where possible. Using computer vision and machine learning, we employ advanced identity detection methods like document forgery detection as well as image and video verification to confirm the authenticity of government-issued IDs and whether they match the individual applying to sell in our store.

But verification is an ongoing activity. Our systems continuously monitor activity throughout a seller’s journey, detecting connections to previously identified bad actors and flagging patterns that may indicate risk. In some cases, unusual activity can trigger reverification. These systems are designed to uncover sophisticated schemes, including bad actors who recruit and pay real people to use their identities to set up seller accounts, which are later sold to others and can serve as fronts for unauthorized activities.

If we discover a bad actor has evaded our verification checks and managed to create a selling account, we swiftly remove them from our store. When we detect bad actors operating in an organized manner, we work with law enforcement globally to pursue legal consequences and dismantle these criminal networks no matter where they operate from.

In addition to vetting who sells in our store, we are doing performance checks as well. Our Account Health Rating (AHR) is one example—a store-level score that reflects a seller’s compliance with Amazon policies and potential risks to customers, legal obligations, and reputation of the store. The score decreases when a seller incurs policy violations and improves through violation resolution and successful order fulfillment. When AHR suspends a seller, it happens for one of two reasons:

1. Poor historical performance: A seller’s score falls below our thresholds due to accumulated violations, whether different types of issues or repeat violations of the same policy.
2. Having a single critical issue: Immediate zero-tolerance violations, primarily involving abuse, compliance issues, or confirmed counterfeit.

This performance-based approach provides clear pathways for sellers to maintain and improve standing, while bad actors face swift consequences that make continued abuse untenable. By combining proactive detection at registration, continuous monitoring and reverification throughout the seller lifecycle, and performance monitoring around non bad actor behaviors, we work to hold bad actors accountable as well as stop unperforming entities before they ever impact customers or our selling partners.

Determining who can sell what in our store: Our approach to selection

Our approach to selection is fundamentally about ensuring customers can always buy with confidence in Amazon’s store.

We make decisions at two levels: what products can be sold, and what requirements must be met to sell them. When determining what to sell, we focus on ensuring products are authentic, safe, and compliant. We may verify supply chains, request safety testing information, or perform other checks to confirm the integrity of products entering our catalog. For products that are more safety sensitive or require special handling, we may require more robust supply chain documentation, verification of storage and handling practices, and specific sourcing channels or minimum inventory volumes, before a seller can list an offer. We also hold a high customer experience bar for product quality, customer service, and delivery performance. Sellers who don’t consistently meet these standards may not be able to list certain products or may lose their selling privileges until they can demonstrate improved performance.

These controls build on a foundational expectation: all products sold in Amazon’s store must be safe and authentic, and compliant with all applicable laws, regulations, and Amazon policies. For well-intending sellers, we’ve developed tools and resources that make compliance more straightforward.

Ensuring product safety and compliance

Beyond authenticity, we ensure products meet rigorous safety standards for our customers. When customers shop in Amazon’s store, they trust that every product they find has been held to the highest safety standards, not just the minimum required by law. And that sometimes means removing selection we believe to be unsafe even if the product meets the prevailing industry standards.

Across the retail industry, some bad actors operate without regard for customer safety. That’s why Amazon goes further to protect customers. We launched our direct product validation program to allow us to work directly with accredited safety laboratories to verify compliance. The program is fully integrated into Amazon’s Account Health Dashboard, allowing accredited providers to validate existing documentation or test products, with results submitted directly from laboratories to Amazon.

We are also exploring new ways to strengthen these efforts. We have begun leveraging new agentic AI that continuously scans for products that may have evaded our existing controls—serving as an audit and furthering our ability to surface potential risks. These agentic audits on top of customer feedback, regulatory developments, and insights from our expert teams, support our constant refinement and evolution of our safety policies.

Protecting intellectual property

Amazon’s automated technology and artificial intelligence (AI) scans billions of attempted changes to product detail pages daily for signs of potential abuse. We deploy advanced AI that identifies intellectual property infringements before product listings go live. Our systems scan text, logos, and images to detect potentially infringing uses of registered trademarks or copyrighted works.

When we receive a valid notice of infringement or customer complaint, our machine learning algorithms use this information to continuously learn and improve our proactive protections. By harnessing multimodal large-language models—systems that can analyze text, visual elements, and pricing patterns simultaneously—we are better equipped to catch even cleverly disguised counterfeits where logos and images are manipulated or obscured to evade detection.

This matters because when customers purchase a branded item, we want them to feel confident they are purchasing a genuine product—and we want brands to know we are protecting their intellectual property.

This proactive approach extends to our work with regulators. Rather than waiting for issues to emerge, we’re initiating conversations with market surveillance authorities and standards organizations to jointly find potential product safety issues faster and prevent them. These collaborations help us stay ahead of emerging safety concerns and ensure our standards align with evolving regulatory requirements.

Our public and private sector collaborations extend beyond regulatory compliance to industry-wide transformation—preparing for changes that will reshape how products are tracked, verified, and protected.

Seeing the full picture: Our approach to innovation

Customer trust isn’t built by solving individual problems in isolation. Threats don’t come from one direction—they require systems that see and use all the information available. And not just to react, but to predict and proactively protect.

That’s why we built systems that analyze details from across the selling and shopping experience—visual elements, textual content, seller identity, behavioral patterns, supply chain data, and network connections. This comprehensive view reveals relationships and patterns that isolated analysis cannot detect.

All of this technology improves thanks to one essential ingredient: feedback. Customers share what they’re experiencing with us in many ways—through reviews, refund requests, and customer service contacts. And we’ve made sure that both customers and selling partners also have direct paths to report concerns when they see something that doesn’t look right: a suspected counterfeit, review abuse, a scam, a potential safety issue, and more. When customers and selling partners tell us what they’re seeing—however they choose to share it—we can act faster and stay ahead of emerging threats.

The future of retail is full of possibility—new technologies, new ways for customers to discover products, new opportunities for businesses to grow. But with every advancement comes new challenges to solve. That’s why we’re committed to inventing on behalf of customers. Every new capability we develop—from AI-based systems that reveal patterns that would otherwise not be discernible to technologies that verify authenticity at unprecedented scale—makes it easier for customers to shop with confidence and for legitimate businesses to thrive. We’re building the foundation for a shopping experience that gets better, safer, and more seamless every day.

AI-powered protection: Stopping fake reviews before customers see them

Since introducing customer reviews in 1995, we’ve accumulated a holistic understanding of authentic review patterns and fraud signals. By combining advanced AI detection and nearly three decades of data, we can identify fake reviews more accurately.

The vast majority of product reviews pass Amazon’s high bar for authenticity and get posted right away. But before a single review appears in our store, artificial intelligence examines thousands of data points simultaneously—account relationships, sign-in patterns, review history, behavioral anomalies. Machine learning models work alongside large language models and deep graph neural networks to detect patterns humans cannot see.

A product accumulating reviews quickly might simply be great. Poor grammar in a review doesn’t necessarily signal fraud. Telling the difference between authentic reviews and abusive or fake ones requires detecting patterns that only become visible through AI-powered analysis of billions of reviews built up over nearly three decades.

However, protecting the integrity of customer reviews isn’t just about technology, it’s about ensuring that the millions of customers and businesses who rely on authentic reviews can trust what they read, every single time.

While AI helps us validate customer feedback about products, we’re also using advanced technology to verify the products themselves—ensuring they meet quality standards before they ever reach customers.

Project PI: Detecting defects before delivery

Inside Amazon’s fulfillment centers across North America, products pass through imaging tunnels designed to sort items based on their destination. Project PI transforms these tunnels into a dual-purpose system—leveraging the images captured to inspect products for potential defects while they move through our fulfillment network. Computer vision technology analyzes the product images to confirm expiration dates haven’t passed and compares actual product images against reference images from the product catalog to detect damage like bent book covers or cracked screens. Our teams then use generative AI to analyze both images and customer feedback simultaneously, enabling faster corrective action.

While Project PI catches what cameras can see, some issues don’t reveal themselves on the surface. They require detection methods that go deeper.

Advanced detection methods can identify sophisticated counterfeits or potential safety concerns at the unit level. But stopping individual bad actors isn’t enough when they’re part of larger criminal operations.

Knowledge graph technology: Understanding the connections that matter

Fraud isn’t always linear. It’s often a cloud of signals that need to be identified, sorted, connected, and analyzed.

Similar to the way social networks map relationships between people, knowledge graph technology helps us connect signals across accounts, listings, transactions, and behaviors—revealing patterns that would remain invisible if we only looked at each data point in isolation. It works alongside machine learning, large language models, computer vision, and expert human investigators to create a comprehensive view of risk. Because protecting customers isn’t just about stopping individual threats—it’s about understanding how those threats connect and operate as networks.

Traditional fraud detection examines signals one at a time: Is this account suspicious? Does this review look fake? But bad actors have learned to game these systems. A review might look authentic on its own, but when you map its relationship to other accounts and behavioral patterns, a network of coordinated abuse emerges.

Many of our systems use knowledge graph technology to map these complex relationships. Deep graph neural networks uncover signals through behavioral patterns, detecting coordinated abuse that would be difficult to detect if each signal were evaluated on its own—whether that’s fake review rings, counterfeit networks, refund fraud schemes, or scam operations.

Knowledge graph technology helps us detect coordinated abuse within our store. But comprehensive protection requires looking beyond—identifying and removing external threats that target consumers around the world.

Stopping bad actors requires changing the fundamental equation that makes counterfeiting, fraud, and organized retail crime profitable in the first place. That requires something no single organization can achieve alone: a coordinated response that spans borders, industries, and jurisdictions.

At Amazon, we take our own data and, in collaboration with brands and law enforcement, identify and dismantle entire production networks at their source. The work of our enforcement teams leads to the interception of shipping containers, raids on illicit manufacturing facilities, and the seizure of entire molds and other machinery used to produce counterfeit goods at scale.

Amazon’s holistic approach to enforcement

Criminal networks don’t operate in isolation. They coordinate across borders, exploit vulnerabilities across the retail supply chain, and fragment their activities to evade detection. If you shut down one operation, they simply restart elsewhere. That’s precisely why our work to pursue bad actors is programmatic, collaborative, and relentless. And why holding them accountable isn’t just about any single consequence. It’s about changing the calculus: making it clear that the risk of operating criminally in or around our store simply isn’t worth it. Because these are industry-wide challenges, a coordinated, holistic response is the only response that works.

A programmatic approach to stopping bad actor networks

Amazon builds the infrastructure to pursue bad actors across every dimension of criminal activity that threatens our store and our customers. This includes:

• Global scam prevention. Impersonation scams—where bad actors pose as Amazon to defraud customers through tech support schemes, phishing, publishing scams, and more—cause harm that extends well beyond financial loss. Amazon pursues these operations through both judicial and non-judicial tools, targeting not just individual bad actors but the infrastructure that enables scams to operate at scale.
• Organized retail crime (ORC). ORC is a coordinated criminal enterprise. Networks systematically target retailers, logistics networks, and retailers through organized refund fraud, stolen goods trafficking, and cargo theft, then attempt to launder those goods back into legitimate commerce. Stolen goods are a serious threat to the entire retail supply chain.
• Cybercrime. Criminal networks target Amazon customer, selling partner, and vendor accounts through account takeover schemes, phishing, and malware—exploiting digital vulnerabilities to steal product, divert financial disbursements, and undermine trust. Amazon’s enforcement strategy targets the full criminal ecosystem: phishing operators, infrastructure providers, and reshipping networks.
• Counterfeit enforcement. Amazon pursues the manufacturers, suppliers, and distributors of counterfeit products through civil litigation, criminal referrals, and coordinated raids.
• Protecting sellers and customers from fraud. This includes pursuing bad actors who exploit Amazon’s returns and refund systems, preventing the theft and misuse of personally identifiable information, and protecting honest selling partners from fake seller registration brokers and account takeover schemes.

Collaboration as a force multiplier

No single company can stop industry-wide criminal networks alone. Amazon’s enforcement approach is built on collaboration—with law enforcement agencies, customs authorities, brands, retailers, and regulators around the world.

Amazon has supported policy solutions and actions by U.S. federal, state, and local law enforcement to curb ORC, participating in over 14 state-run ORC task forces where we work regularly with industry and law enforcement. Our teams provide on-the-ground support during operations, testimony at trial, and proactive intelligence sharing through channels like the EUIPO IP Enforcement Portal.

This model extends across every enforcement area: Amazon became a founding member of the National Elder Fraud Coordination Center to combat fraud targeting older adults; coordinated operations with European law enforcement resulting in arrests of phishers; and worked with law enforcement across North America and Europe to dismantle cargo theft networks responsible for hundreds of millions in losses.

And in 2026, Amazon is expanding Amazon’s Counterfeit Crimes Unit’s (CCU) enforcement efforts to India through the creation of a dedicated team consisting of attorneys, investigators, and data analysts working to protect customers and selling partners from counterfeits and criminal organizations that manufacture and distribute counterfeit goods. Leveraging Amazon’s existing investigatory and legal resources, the team will build actionable referrals using internal data while establishing critical relationships with Indian law enforcement, prosecutors, and brands to pursue bad actors through both criminal referrals and civil litigation. This expansion builds on the proven success of the CCU’s dedicated team in China.

The goal isn’t just to stop bad actors from operating in our store. It’s to stop them from operating anywhere.

The results of this work

Enforcement actions result in arrests, criminal convictions, asset seizures, civil judgments, and restitution for victims. Every investigation, every case, every enforcement action makes the next one more effective. Each operation helps us understand criminal networks better, spot patterns earlier, and stop counterfeits, scams, and fraud before they reach customers anywhere.

When retailers share intelligence with law enforcement, when brands collaborate with customs authorities, when governments establish frameworks that enable swift action—bad actors are held accountable everywhere they operate, not just in isolated regions. The cases below demonstrate why this collaborative infrastructure matters.

One of the most powerful defenses against fraud, counterfeiting, and unsafe products is an informed customer. At Amazon, we believe consumer education isn’t just about sharing information—it’s about delivering the right knowledge at the moment people need it most. Whether that’s safety guidance arriving hours after a natural disaster strikes, scam prevention tips during peak shopping periods, or resources that help new parents navigate overwhelming choices, our approach centers on meeting customers where they are with information they can immediately use.

But we also recognize that building trust requires looking beyond our own store. The challenges consumers face—from sophisticated phishing schemes to counterfeit goods—don’t respect boundaries between the public and private sector. That’s why our education initiatives include collaborations with governments, safety organizations, educational institutions, and advocacy groups around the world. Together, we’re working to ensure that every consumer, has access to the knowledge and tools they need to shop confidently and safely—not just in Amazon’s store, but anywhere they shop.

The initiatives below demonstrate how this vision comes to life across different regions, populations, and challenges—each one designed to empower consumers with knowledge that protects them and builds lasting trust.

These initiatives represent more than individual programs—they’re building blocks of a trustworthy shopping experience for everyone. Every workshop, safety guide, and collaboration creates ripples that extend far beyond a single transaction. As bad actors grow more sophisticated, so does the need for a collective response. And as commerce expands into new communities, consumer education expands alongside it—meeting people where they are, in languages they speak, through channels they trust.

Behind every statistic in this report are real people—customers discovering products that bring joy to their families, small businesses growing with confidence, brands launching innovations they’ve worked years to create, and teams collaborating with industry leaders around the world.

Amazon’s mission is to make customers’ lives better and easier every day by relentlessly inventing on their behalf. Maintaining a trustworthy shopping experience is part of that mission—a journey of constant improvement and innovation. The systems in this report aren’t finished. They’re evolving every day.

The challenges facing global retail today didn’t exist a decade ago. Neither did many of the solutions. As retail evolves, so must the infrastructure that protects it. We’re working to build that future—not alone, but alongside brands, law enforcement, regulators, and industry collaborators around the world. Together, we’re inventing new ways to detect threats before they emerge, dismantle criminal networks at their source, and empower consumers with knowledge that protects them everywhere they shop.

What you’ve seen in this report represents the most comprehensive, collaborative approach to trustworthy shopping in the industry—and it’s the foundation we build on every day. We’ll keep pushing: systems that get smarter, collaborations that reach farther, and protections that work harder. Because earning customer trust isn’t a destination. It’s the work.