BrandShield vs Red Points: how they compare in 2026

The models do the finding. AI.ClusterX, BrandShield's own predictive clustering engine, reads infrastructure, hosting patterns, content signals and behavioral indicators to judge whether a domain or asset is malicious and to tie it to known threat activity. Around it sit agentic AI, language processing, reverse image search for altered logos and product shots, and AI prioritization. AI Platforms Protection goes a step further and uses prompts to test what chatbots tell shoppers about a brand. Remove the models and a business still remains. A 24/7 security operations center, named enforcement managers, standing relationships with registrars, hosts and marketplaces, and a case and evidence workflow would still sell takedowns. So the AI drives detection on top of an enforcement service rather than being the whole product.

The models are the engine of detection, and an enforcement business sits under them. Red Points describes five layers. Autonomous agents run prioritization, tagging, enrichment, rule suggestions, documentation and reporting. A governed tier of fixed, auditable logic handles validation, enforcement and evidence. Engineers placed with each customer tune the system, a brand knowledge layer holds ten years of playbooks, and 2.7 billion data points a month feed it. Vision AI, custom models for each product and the Seller Risk Score are model work and would not survive without the models. The enforcement network would. More than 300 removal agreements, direct API connections with named platforms, priority reporting paths, the IP document store and the case workflow would all keep running.

The product cites no legal authority, so grounding in primary law and a citator do not apply. What matters is whether a detection rests on evidence a reviewer can see, and whether anyone measures how often it is right. Grounding is described. BrandShield states what AI.ClusterX reads to reach a verdict on a domain or asset, says evidence is collected automatically, and builds each case ready for takedown, so the reviewer sees the material behind each call. Threats found in chatbot answers are logged with documentation for brand protection, legal and marketing teams. No precision, recall, false positive rate or test set is published. The figures BrandShield does give, 98% of threats detected and removed within the agreed service level and a 98% takedown success rate, measure removal rather than whether each detection was right.

The product cites no legal authority, so grounding and accuracy disclosure are the questions. Grounding is documented. Every decision is logged with what was detected, why, what triggered it and what happened next. Vision AI shows the logos, patterns or products that triggered a match. Evidence is generated and formatted to each infringement type, platform and jurisdiction, so a reviewer sees the triggering item itself. No measurement is published, with no precision, recall, false positive rate, test set or outside validation. Clause 8.9 of the service terms provides the output as is, with no warranty that it is accurate. The figures Red Points gives are economic and enforcement figures, such as a cost per enforcement 10.3 times lower than analyst based models and enforcement success rates at named customers.

The modes are published, and so is the review step before anything is removed. Customers choose fully managed, self service or hybrid delivery, which sets how much the system and BrandShield's staff do alone. The entry product automates detection, investigation and, if the customer wants, response, with self managed remediation. Under section 5.1 of the terms, BrandShield proposes suggested targets and starts a takedown only after the customer confirms them, or follows a policy the customer has set. Analysts in the 24/7 security operations center review what the AI flags before enforcement, and named enforcement managers work complex cases with registrars and hosts. What is missing is what happens when a takedown was wrong. No correction, appeal or withdrawal route for a wrongly removed asset is published.

Red Points publishes what runs alone, what constrains it and how a wrong takedown comes back to a person. Autonomous agents handle prioritization, tagging, enrichment, rule suggestions, documentation and reporting, and escalate to a person when required. Validation, enforcement and evidence are reserved for a separate governed tier of fixed, auditable logic, so the agents do not make the enforcement call. On its impersonation listing, Red Points says customers set automation anywhere from fully automated takedowns to manual confirmation. Its notice to reported sellers says takedowns are filed once the client validates a violation. Every decision is logged with what was detected, why and what happened next. When a reported seller sends a counter notice, clause 8.8 of the service terms has Red Points tell the client in writing and withdraw the takedown if no instruction arrives within two business days. The client can also have a takedown withdrawn under clause 8.7, though restoring the content is up to the platform.

The July 2026 Frankie4 story reports a 97.5% takedown rate for scam websites and 100% removal of fraudulent ads for the Australian footwear label, with no period or basis given. The la Vie en Rose story reports 100% takedown of fake ads. The August 2026 Scrub Daddy story quotes its compliance coordinator, John Candita, and dates the work from December 2024, with results told in words rather than figures. Legence Bank's Drew Morris says a fake site was offline within 48 hours, and Liontrust reports a 100% takedown success rate on external threats. Dropbox, LendingClub, Playtika, eToro, Bang & Olufsen and Ferrara also appear by name. BrandShield says more than 500 brands use the platform. Its figure of 98% of threats removed within the agreed service level has no stated basis.

Customers are named and tied to figures, but the figures carry no method. The ASICS story quotes its Global Brand Protection Manager and reports more than 200,000 counterfeit listings identified in the first four months. It says listings enforced each month rose from 150 to more than 20,000, and gives a 99.7% enforcement success rate and $2.5 billion in counterfeit products removed. Purple's James Larson, VP and Deputy General Counsel of IP, is tied to more than 4,700 enforcements a month and a 94.5% success rate. Topcon's Satoru Ishiguro is tied to $22 million in infringements removed, FILA Italy to 638,653 listings removed, and Cotopaxi to 4,700 fake websites removed in one month and a 95% automation rate. Kaspersky, Burton, KEEN and Essity also appear with figures. Nothing explains how dollar values were set or how a success rate is counted.

Confidentiality is promised in general terms, and the rest of the picture is missing. Section 10 of the terms is a mutual clause. It covers all non public data and information either side discloses and bars disclosure to third parties without written approval. It allows disclosure required by statute, regulation or court order, with prompt notice to the other side and help to limit it. The clause restricts disclosure and says nothing about how the receiving side may use the information. No position on model training is published. Nothing addresses separation between brands or users, retention or deletion of customer material, or privilege and work product. That matters because a customer sends its trademark portfolio, its enforcement instructions and the letters of authorization needed to act in its name. The data processing agreement applies on subscribing and is not public.

Red Points' service terms make real commitments on confidentiality and tell the client plainly that privilege does not attach. Clause 1.5 says Red Points is not a law firm, and that communications with it are protected by the agreement and its privacy policy but not by attorney client privilege or work product. Clause 5 binds each side to keep the other's non public information confidential for the contract and two years after, or as long as a trade secret lasts. It limits use to what the agreement needs and requires return or destruction on request at the end. Clause 4.3 licenses the client's material to Red Points only as needed to provide the services. Platform data is deleted thirty days after termination. What is missing is separation between brands or users inside an account, and a stated position on model training. Red Points' own pages describe custom models built for each customer's products.

BrandShield's terms draw one line between its service and legal work, and stop there. Section 5 makes the Takedown Service a separate opt in service that proposes targets, acts after the customer confirms, and may need letters of authorization to act for the brand owner. If a takedown does not stop the infringer, the customer gets a report and can pursue legal action separately, such as dispute proceedings, court action or a local attorney, outside the scope of the service. Even so, the service itself sends cease and desist letters, ICANN and registrar reports and search removal requests in the customer's name. BrandShield also cites more than a decade of IP legal and enforcement experience. No statement says BrandShield is not a law firm, gives no legal advice or forms no attorney client relationship. No jurisdiction limits are named, though enforcement runs across registrars and platforms worldwide.

Red Points states plainly, in its service terms, what it is and is not. Clause 1.5 says it is not a law firm, that its employees cannot give legal guidance, and that the services are not a substitute for an attorney. It is barred from any advice, opinion or recommendation on legal rights, remedies or strategies, and the client makes every decision on its rights. Legal questions go to the client's own counsel, and no attorney client relationship arises. Clause 8.6 has the client warrant that qualified people review and verify what it sends, and that it reports infringement only on a good faith belief in the claim. The website legal notice repeats the same position. What is not addressed is the competence and supervision duty of the lawyer who often directs the program, or any limit by jurisdiction.

No governance position is published for a system that decides whether a third party's domain, listing or profile is malicious and then acts to remove it. BrandShield has no trust center, responsible AI page or governance page, and its legal footer holds only the privacy policy, the terms of use and a cookie policy. No accountable owner is named. Nothing describes testing before release or publishes evaluation results. Nothing addresses whether output is uneven across sectors, languages or countries. That gap matters on this product, because a false positive falls on a third party seller or site operator rather than on the buyer. A July 2026 BrandShield post says human analysts check AI findings before enforcement. That is a review step in daily operations, not a governance framework with an owner, a test regime or published results.

What Red Points publishes is an architecture and a compliance claim, not a governance position. Its governed tier is described as fixed, fully auditable logic, and every decision is logged with what was detected, why and what triggered it. Its company information page calls Red Points compliant with the GDPR and the EU AI Act, and describes expert review in the loop. No accountable owner is named. No testing regime, evaluation result, AI policy or model card is published, and nothing explains what the EU AI Act claim rests on. Nothing addresses uneven output, although the system makes decisions about third parties. A validated detection removes a named seller's listing, and a Seller Risk Score labels a trading entity as a repeat infringer. No error rate, appeal figure or analysis across seller types, categories or countries is disclosed.

BrandShield publishes a website privacy policy and general terms, and neither says what happens to platform data. The privacy policy, updated July 2026, names the platform and products in its opening definition and then says it governs only information collected on the website. It covers browsing, registration, cookies, ad servers, promotional messages and payment details, and offers deletion of personal information on written request. The terms add little for platform data. On termination, access to work in progress and hosted content is disabled, and an account inactive for 180 days may be closed. No retention period for detections, evidence or case records is stated. No deletion commitment on termination, access control description, subprocessor list or breach notice practice is published. A data processing agreement applies once a customer subscribes, and it is not public. The only security statement is that BrandShield takes precautions, with an express caveat that its systems do not provide absolute security.

A published data processing agreement, dated April 2022, binds Red Points to process personal data only on the client's documented instructions. Schedule 3 names two subprocessors, AWS for hosting in Ireland and Salesforce for customer records in the United States. Red Points must give twenty days' notice of a change, and the client may object. Schedule 2 lists access control, password rules, logging, encryption, availability controls and firewalls. A confirmed security breach is reported to the client within 48 hours. Personal data is deleted on termination. The service terms add that the client can download its platform data for thirty days after the contract ends, after which Red Points deletes it. Clause 2.3 commits Red Points to security consistent with good industry standards. Nothing states how long detections, evidence and uploaded brand assets are kept during the contract.

BrandShield's liability position is a standard limitation clause, and it leaves the product's main exposure with the customer. Section 16 of the terms caps BrandShield's total liability at the fees paid or due in the twelve months before the claim. It excludes indirect and consequential loss and lifts the cap only for gross negligence or willful misconduct. Section 15 provides the service as is, with no warranty that it is free of errors or that its information is accurate. BrandShield gives the customer no indemnity. The customer indemnifies BrandShield under section 17.1, including for takedown actions taken on the customer's instructions or policy. That matters because a wrong detection removes a third party's listing, domain or account, and the cease and desist letters and platform reports go out in the customer's name. No insurance position, warranty on detections or remedy for a wrongly removed asset is published. Fees cannot be canceled or refunded under section 6.1.

Clause 8.5 of Red Points' service terms warrants that the services perform substantially to their specifications and that professional services are done in a workmanlike manner. Clause 9.1 defends and indemnifies the client against third party claims that the services infringe a patent, copyright, trademark or trade secret, with replacement, modification or a prorated refund as remedies and named exclusions. Clause 10.1 caps Red Points' total liability at the fees paid in the prior twelve months, and clause 10.2 excludes indirect and consequential loss. The output itself is provided as is under clause 8.9, with no warranty that it is accurate or reliable. The client indemnifies Red Points under clause 9.2 for claims arising from takedowns it instructed or authorized. For shutting down a website on the client's instruction, Red Points is liable only for gross negligence. No insurance position is published.

What BrandShield publishes is a buying channel rather than a documented integration. The AWS Marketplace listing lets a customer buy and run the service inside an existing AWS account and pay through the AWS bill. Online Brand Protection and External Cybersecurity are presented as one connected platform. Everything else described as coverage is monitoring reach. Marketplaces, social platforms, app stores, domains, paid ads and the dark web are watched, and registrars and hosts are counterparties for enforcement rather than systems the customer connects. No API or developer documentation is published, and nothing says what data moves in which direction or what a customer must set up. The terms name one piece of customer side setup, a Host Verifier component embedded for the Website Duplication Detection Service to raise alerts from unfamiliar sources. No connection to a trademark management, docketing or document management system is described, which is where an IP counsel's own records sit.

Real integrations are named and described, short of documented depth. Direct API connections with Meta, Google and Bing are described as letting enforcement act at once and get past geographic limits. More than 300 removal agreements and platform partnerships cover more than 5,000 platforms, among them Amazon, Alibaba, eBay, Taobao, Walmart, Instagram, TikTok, Douyin, WeChat and Pinduoduo. Customers log in to a dedicated application, Document Management holds the IP rights being enforced, and the service can be bought through AWS Marketplace. At the end of a contract, platform data can be downloaded as a CSV file for thirty days. No developer or API documentation is published, and nothing describes what syncs in which direction or what a customer must configure. The platform relationships are escalation paths rather than integrations an implementer could use. No connection to an IP portfolio, docketing or trademark management system is described.

Software as a service is the delivery model, and neither the tenancy model nor the region is stated. The AWS Marketplace listing records the product as deployed on AWS, and BrandShield presents that as a way for security teams to run domain and phishing protection inside their existing AWS account. Nothing else is published on where the service runs. There is no statement of shared or dedicated tenancy, no list or choice of regions, and no split between where data is processed and where it is stored. The terms say Standard Contractual Clauses may apply to personal data from Europe, Switzerland or the United Kingdom under the data processing agreement, which is not public. That is the one transfer mechanism named, for a company contracting from Israel with customers in Europe and the United States.

Software as a service on AWS is the delivery model, with partial residency detail. Red Points' AWS Marketplace listings record the product as deployed on AWS. The data processing agreement names AWS as the hosting subprocessor in Ireland, and Salesforce in the United States for customer records under Standard Contractual Clauses. It says Red Points prefers subprocessors inside the European Economic Area and uses an approved transfer mechanism outside it. Nothing states whether tenancy is shared or dedicated, which other regions are offered, or where the models process data as distinct from where it is stored. The service terms are governed by Spanish law, with disputes heard in the courts of Barcelona.

No independent security attestation is published. BrandShield has no trust center or security page, and its footer carries no certification marks. The recognition it cites, from G2, the Cybersecurity Breakthrough Awards and Frost & Sullivan, is not an attestation. The terms carry no security commitment from BrandShield, while section 4.1 obliges the customer to report a suspected breach at once. The privacy policy says payment card data is encrypted in transit and that BrandShield takes precautions, with an express caveat that its systems do not provide absolute security.

Marks appear on the site with no scope, date or report. Every page footer shows a TÜV certificate image and an AWS qualified software mark, with no text naming a standard, scope or period, so what the TÜV mark certifies is not stated. The footer also shows Google Trusted Copyright Removal Program membership, Google Trusted Reporter and YouTube Certified Multi Channel Network status. Those are program memberships, not security attestations. There is no trust center. The Security Policy page states four principles and gives a Security Manager the job of analyzing risk, with no standard, auditor or report. The data processing agreement lists security measures and a 48 hour breach notice, and names no certification.

Only the in house engine is named, not what sits under it. AI.ClusterX is described as a proprietary predictive clustering engine. Its inputs are infrastructure, hosting patterns, content signals and behavioral indicators, used to decide whether an asset is malicious and how it connects to known threat activity. The product pages add agentic AI, language processing and reverse image search as techniques. No model is named and no model provider is identified. Nothing says where inference runs, and no commitment to tell customers when any of this changes is published. The AWS deployment on the marketplace listing describes the hosting, not the models.

Proprietary models are referred to throughout, without naming what sits under them. Red Points' pages describe Vision AI models that scan logo, shape, packaging, text, color and known fakes at once, and brand specific models that rank each match by whether a consumer would be deceived. They add custom models for each product, predictive models trained on 2.7 billion data points a month, the Seller Risk Score, an AI Copilot and language processing. No model or provider is named. Nothing describes the architecture beyond the visual matching, and no commitment to tell customers when models change is published. The data processing agreement names AWS and Salesforce as subprocessors and no model provider.

Real pricing is published for part of the range. BrandShield is the seller of record on its own AWS Marketplace listing, with its own copy and license agreement. The BrandShield Managed Anti-Phishing Solution, the packaging of BrandShield Resolve, costs $22,000 for a twelve month contract. That covers unlimited standard takedowns, a capped allowance of managed premium takedowns and 24/7 security operations support. It is billed by units on one dimension, and the price scales with units committed rather than with threat volume or takedown count. Extra AWS infrastructure costs may apply, and fees are nonrefundable. The terms add that subscriptions renew automatically unless either side gives sixty days' written notice. The two enterprise products, Online Brand Protection and External Cybersecurity, carry no published price. BrandShield's own pages offer a demo rather than a price.

Two prices are published and the structure is described in depth. On AWS Marketplace, where Red Points is the seller of record with its own service terms, a twelve month Professional Program for three trademarks with marketplace, social media and image protection costs $39,000. The Digital Impersonation Removal Professional Plan, covering social media, websites, domains and apps, costs $64,000 for twelve months. Both offer a 24 month term at a stated discount. The pricing page sets out five programs with named tiers and a feature comparison covering trademark counts, channels, managed service, dashboards and more than ten add ons. Each is a flat fee for an agreed scope, with unlimited detections, takedowns and users inside it. Every tier on Red Points' own site routes to a pricing request, so Starter and Enterprise carry no figure. The service terms add that fees are paid in advance, that each renewal brings the increase set in the order, and that late fees carry interest at 1.5% a month.

Coverage is segmented precisely and the boundaries are left open. BrandShield publishes pages for twelve industries, among them pharmaceuticals, financial services, fashion, gaming and luxury, with separate pages for small businesses and enterprises. Its use case pages cover job scams, social engineering, piracy, grey market goods and AI search. Six buying roles have their own pages, from founders and legal teams to compliance, IT, threat intelligence and information security. A legal teams page beside a CISO page tells a buyer that both are served. Resolve is aimed at lean security teams. What is not stated is where coverage stops, such as the countries, languages or marketplaces the service does not reach.

Coverage is described in detail, and its limits are only partly stated. Five program families are published, with channels named across marketplaces, social media, search, ads, apps, domains, video and AI shopping. Regions are North America, EMEA, APAC, LATAM and mainland China, with Chinese platforms listed one by one, and Red Points says it serves brands in more than 100 countries. Its reports are written for legal, brand protection, ecommerce, security and executive teams, and Red Points says it works with mid market and enterprise companies. Red Points says it is not a law firm, not offline enforcement and not a consumer research tool. Nothing states which infringement types, platforms or jurisdictions fall outside the service. Clause 8.7 of the service terms notes that some content cannot be removed because of a platform's own policies or gaps in the client's registrations.