Perly Consulting │ Beck Eco

The State of Play

A living index of AI adoption across industries — where established practice meets the bleeding edge
UPDATED DAILY

The AI landscape doesn't move in one direction — it lurches. Some techniques leap from experiment to table stakes in a single quarter; others stall against regulatory walls, technical ceilings, or organisational inertia that no amount of hype can dislodge. Knowing which is which is the hard part. The State of Play cuts through the noise with a rigorously maintained index of AI techniques across every major business domain — classified by maturity, evidenced by real-world adoption, and updated daily so you always know where you stand relative to the field. Stop guessing. Start knowing.

The Daily Dispatch

A daily newsletter distilling the past two weeks of movement in a domain or two — delivered to your inbox while the index updates in the background.

AI Maturity by Domain

Each dot marks the weighted maturity of practices within a domain — hover for a brief summary, click for more detail

DOMAIN
BLEEDING EDGEESTABLISHED

Intellectual property monitoring & enforcement

LEADING EDGE

TRAJECTORY

Stalled

AI that monitors for intellectual property infringement including trademark violations, patent infringement, and counterfeit goods. Includes marketplace monitoring and image-based infringement detection; distinct from patent analysis which researches the landscape rather than monitoring for violations.

OVERVIEW

AI-powered IP enforcement has reached production maturity and crossed into mainstream adoption while remaining concentrated in specialist brand protection teams — the defining tension sustaining a leading-edge plateau. Specialized vendors process 70+ million marketplace items daily with detection accuracy above 94%, and platforms now process billions of data points annually ($3.7B in physical goods authenticated, 15M marketplace counterfeits removed by major platforms). Yet organisational readiness gaps persist: enforcement at scale requires multi-layer human judgment, false positive management at industrial scale erodes business trust, and jurisdictional fragmentation across 118+ countries constrains enforcement coordination. The practice exhibits sustained production deployment longevity (six years at leading-edge tier), measurable customer outcomes (cost reduction, takedown velocity), and vendor ecosystem maturity — but structural barriers (platform cooperation friction, false positive cost, legal judgment requirements) block mainstream adoption beyond specialist teams.

The accelerating threat (impersonation surged 460% in 2025, counterfeiting now 2.5% of global trade) has driven market inflection: 90% of surveyed companies now face AI-accelerated threats, 82% plan to increase investment, and government enforcement (US, China, Europe, Canada, Australia) has formalized AI deployment at scale. Yet deployment velocity remains constrained by the enforcement paradox: AI excels at detection but cannot enforce without registrar relationships, legal escalation, and human expert judgment.

CURRENT LANDSCAPE

April 2026 data confirms production-scale deployment maturity across major enterprise platforms and government agencies. Amazon's Omniscan system removed 15 million counterfeits in 2025 with coordinated law enforcement, pursuing 32,000 bad actors across 14 countries and blocking anticipated threats 8 days before brand notification — demonstrating large-scale IP enforcement integrated with criminal justice. MarqVision deployed Marq AI (multi-agent enforcement engine) achieving 99.8% accuracy with median 5.3-hour response for domains (37x faster than industry 2-8 day baseline) and 11.3 hours for paid ads across 48,253 real incidents; company raised $48M Series B (April 2026, total $90M) targeting $100M ARR by mid-2027 with 350+ brand customers. The USPTO deployed three AI tools for trademark examination: image-search for visual similarity detection, mark description generator, and Class ACT for automated classification (immediate processing vs months historical), signaling government-scale institutional adoption. China's State Council reported 37,000 patent/trademark cases and 38,000 customs seizures (75.75 million items) in 2025 with 2026 priorities emphasizing emerging fields and e-commerce enforcement. Entrupy's physical authentication platform processed $3.7B worth of goods in 2025 (33% YoY growth) with 99.86% accuracy, demonstrating production-scale counterfeit detection in secondary markets.

Market adoption signals have strengthened: 90% of surveyed B2C companies now face AI-accelerated threats with 78% experiencing 5%+ revenue loss; 82% plan increased investment. Yet critical structural barriers remain entrenched. False positive management at industrial scale erodes business trust — generative AI enables harder-to-detect impersonation (synthetic sellers, AI-generated images), and model drift increases false positives without continuous retraining. The enforcement paradox persists: AI excels at detection (billions of datapoints processed daily) but cannot enforce without human expert judgment, registrar relationships, legal escalation, and abuse desk responsiveness. SunTec's analysis documents that enforcement decisions require contextual knowledge (distinguishing counterfeits from authorized resellers, gray markets, legitimate reuse), and false positive consequences range from operational drag to legal liability and channel partner friction. The practice remains at production scale and leading-edge tier, sustained by six-year continuous deployment longevity and vendor ecosystem maturity, but organizational readiness gaps (specialist-team concentration), jurisdictional fragmentation (118+ countries), and the enforcement judgment requirement continue blocking mainstream adoption beyond large enterprise specialists.

TIER HISTORY

ResearchJan-2018 → Jan-2018
Bleeding EdgeJan-2018 → Jan-2020
Leading EdgeJan-2020 → present

EVIDENCE (103)

— Official China enforcement metrics from April 2026 press conference: 37K patent/trademark cases, 38K customs seizures (75.75M items), 1.27M calculated trademark rejections, 3.35K ex officio invalidations; 2026 priorities emphasize emerging fields, e-commerce, livestream commerce enforcement.

— Entrupy's AI product authentication platform processed $3.7B worth of luxury goods in 2025 (33% YoY growth), achieving 99.86% accuracy with <60 second authentication; $3.34B authenticated, $294M disputed; demonstrates production-scale counterfeit detection in secondary market.

— Amazon's Omniscan system (ML-powered counterfeit detection) deployed across US, Canada, UK, Turkey, Saudi Arabia, Europe; removed 15M counterfeits in 2025, pursued 32K bad actors across 14 countries, blocked 8-day-ahead phishing attack; demonstrates production-scale IP enforcement with law enforcement coordination.

— MarqVision deployed Marq AI (multi-agent enforcement engine) targeting fake domains, paid ads, website content; 48,253 real incidents: 99.8% accuracy, 5.3h median domain takedown (37x faster than 2-8 day industry baseline), 11.3h for paid ads; human-in-the-loop review embedded in pipeline.

— SunTec analysis documents gap between detection speed and enforcement accuracy: false positives erode business trust, create operational drag, pose legal liability; generative AI increases harder-to-detect impersonation (synthetic sellers, AI-generated images, AI-written listings); model drift causes false positives to increase without retraining.

— USPTO deployed three AI tools for trademark monitoring/examination: image-search (visual similarity reverse search), mark description generator (live 2026-04-23), Class ACT (automated classification achieving immediate processing vs months historical); government-level institutional adoption in IP infrastructure.

— Peer-reviewed research (Computer Fraud & Security, April 2026) on multimodal ML architectures for IP enforcement; evolution from text-based to multimodal (textual, visual, behavioral); contextual intelligence enables distinction between intentional infringement and honest policy breaches; emphasizes real-time prevention and cross-platform monitoring ROI.

— Critical assessment from SunTec India: automated detection excels at scale but enforcement decisions (takedowns, C&D, escalation) require human judgment to avoid false positives damaging channel relationships; flags legitimate resellers, gray markets, authorized reuse scenarios requiring expert review.

HISTORY

  • 2018: IP monitoring platforms consolidated around AI image recognition and marketplace automation. Red Points reached 300 customers with 350% annual revenue growth. Clarivate acquired TrademarkVision for visual trademark detection. MarkMonitor adapted to GDPR restrictions. Global counterfeit goods cost estimated at $1.2 trillion annually.
  • 2019: Red Points scaled to 550+ customers with 500k+ detected incidents monthly (100%+ YoY growth). USPTO deployed AI to detect fraudulent filings amid 10-fold surge in Chinese applications. WIPO and USPTO launched public consultations on AI's role in IP enforcement. Accenture filed patents for AI counterfeit detection. IP practitioners showed high awareness (83%) but low implementation (20% using AI), indicating maturation of vendor solutions but slower enterprise adoption.
  • 2020: EU-funded Microguard project achieved 98% accuracy in counterfeit pharmaceutical detection with 115,000 labels deployed for Incepta Pharma. MarqVision's platform blocked 1,000+ counterfeit listings across ASEAN e-commerce sites. WIPO convened 2,000+ participants from 130 countries for a second policy session on AI and IP enforcement. Industry adoption momentum continued: INTA confirmed AI use by IP registries was "gaining prevalence." The practice remained constrained by false positive rates, jurisdictional fragmentation, and platform cooperation challenges despite growing institutional recognition.
  • 2021: Venture capital and institutional funding accelerated: SoftBank Ventures backed MarqVision (founded by Harvard Law graduates) with $5M seed capital as its first legal-tech investment; Red Points secured €15M from the European Investment Bank for expansion. Major jurisdictions formalized AI policy engagement: UK IPO and INTA both published frameworks addressing AI in trademark clearance and enforcement, with 92 UK stakeholders supporting liability rules for AI infringement. Academic institutions (Cambridge, Max Planck) published research on AI in IP border enforcement and holistic regulatory approaches. MarqVision achieved $1M ARR with 50% monthly growth. The category remained supply-constrained—while detection technologies matured, market adoption continued to lag awareness due to jurisdictional complexity and platform cooperation friction.
  • 2022-H1: Specialized vendors expanded into emerging attack surfaces: MarqVision launched NFT counterfeit detection (March 2022) across OpenSea and other NFT markets, while Entrupy's physical authentication device reached 65+ countries with >99% accuracy for luxury goods. Institutional adoption signals strengthened: UK government published dedicated counter-infringement strategy (Feb 2022) committing to technology adoption and intermediary engagement. Yet regulatory caution persisted: UK government response (June 2022) noted AI enforcement remained "early stage" with limited evidence for policy change. Real-world deployments demonstrated measurable outcomes—MarqVision's logo detection achieved 87% efficiency gains in counterfeit removal—but coordination bottlenecks (platform cooperation, jurisdictional fragmentation, false positive rates) remained endemic constraints on enforcement velocity and scale.
  • 2022-H2: Platform consolidation and institutional adoption accelerated. MarqVision removed $1M in counterfeit Ralph Lauren merchandise with 97% accuracy across 1,500+ retailers; Red Points secured €20M funding and reported €1.7B in counterfeit value removed in 12 months while serving 1,000+ brands. Government agencies globally formalized AI deployment: USPTO, IP Australia, Canadian IPO, Norwegian IPO, and EUIPO all adopted machine learning tools for trademark examination and bad faith detection. Academic research validated physical counterfeit detection (96.55% accuracy using OCT-ML for banknotes) and quantified measurement gaps in enterprise brand protection ROI assessment across industries. Venture ecosystem confidence remained strong despite regulatory caution about enforcement maturity.
  • 2023-H1: Red Points and MarqVision demonstrated production-scale deployment. Red Points removed 10,100+ counterfeits across 700+ marketplaces for Highwave with 93% accuracy. MarqVision's survey (295 companies) revealed 68% of DTC brands counterfeited but only 29% had documented enforcement strategies, indicating high market opportunity despite awareness gaps.
  • 2024-Q1: MarqVision launched Marq AI generative AI suite with 99%+ accuracy and 180x efficiency gains, signaling evolution toward agentic IP enforcement. Market reached USD 2.99B with 12.8% CAGR growth projection. Adoption barriers persisted: 63% of legal professionals hesitant about GenAI deployment due to data security and IP risks. EU AI Act (March 2024) introduced regulatory constraints. Academic research highlighted deepfakes as new enforcement challenge vector.
  • 2024-Q2: MarqVision demonstrated real-world SMB deployment protecting Darn Tough Vermont from impersonation scams, automating detection of hidden fake websites. Academic research (Hanyang Law Review, 4iP Council) explored AI's role in trademark similarity assessment for litigation and broader IP enforcement use cases. Vendor analysis identified counterfeiters' adoption of AI evasion tactics as key constraint on enforcement velocity.
  • 2024-Q3: Red Points reported processing 30M marketplace links daily across 1,300+ brand clients; Entrupy expanded AI authentication capabilities in luxury resale market. Critical assessment of trademark monitoring tools revealed persistent data quality issues (update delays, missing records, poor translations) limiting reliability. Clarivate survey showed 83% of IP professionals with non-core responsibilities lacked AI deployment despite rising awareness. Counterfeit goods valued at $2.8T globally with $600B annual brand losses continued driving vendor innovation but constrained by platform cooperation and jurisdictional fragmentation barriers.
  • 2024-Q4: Academic research advanced counterfeit detection via smartphone imagery (99.71% accuracy on branded garments using deep neural networks), while TaskUs and Red Points announced strategic partnership combining AI detection with human enforcement, with Red Points serving 1,300+ brands and protecting $500M in discovered infringements. New academic projects demonstrated continued innovation in ML-based trademark infringement detection and automated enforcement workflows. However, broader adoption remained constrained by organizational readiness gaps and the fundamental structural challenges—jurisdictional fragmentation, platform cooperation friction, data quality limitations, and persistent ROI measurement gaps—that had blocked mainstream tier adoption for multiple years despite leading-edge technical capabilities.
  • 2025-Q1: Enterprise vendor maturation accelerated with Corsearch TrademarkNow achieving GA deployment at Fortune 500 scale with 40% efficiency gains in trademark clearance. Market analysis documented 80M+ active trademarks globally and 65% of luxury brands facing counterfeits on major platforms. ML capabilities advanced to 94.3% accuracy in unauthorized detection with sub-100ms processing latency, signaling technical readiness for enterprise scale. However, adoption barriers remained entrenched: organizational deployment concentrated in specialist brand protection rather than general IP departments; counterfeit goods market at $2T annually with counterfeiters adopting AI evasion tactics; EU DSA driving 24-hour removal mandates creating new compliance friction; and ROI measurement gaps persisting as adoption friction point. Structural bottlenecks—jurisdictional fragmentation, platform cooperation constraints, tool reliability limitations, and organizational readiness—remained largely unchanged, perpetuating the supply-demand imbalance that had defined the practice's leading-edge plateau.
  • 2025-Q2: Technical maturity deepened with peer-reviewed research confirming e-commerce accuracy of 94% and social media 87%; Clarivate launched RiskMark GA with generative and predictive AI for trademark similarity assessment. Luxury goods deployments accelerated: Lacoste, Prada, and Louis Vuitton deployed AI and blockchain authentication systems. Brand protection software market projected 18.94% CAGR to $3.172B by 2032. However, critical practitioner assessments documented persistent limitations: AI struggles with contextual distinction between counterfeits and parallel goods, faces GDPR/privacy constraints, and lacks clear accountability mechanisms. Counterfeiting escalated to 2.5% of global trade with counterfeiters adopting AI evasion tactics. The practice remained at leading-edge tier: technical capabilities matured, vendor ecosystems solidified, but enforcement barriers (jurisdictional fragmentation, platform cooperation friction, tool reliability limitations, organizational readiness gaps) remained unbroken, sustaining the supply-demand plateau.
  • 2025-Q3: Vendor ecosystem continued scaling: MarqVision completed $48M Series B funding (total $90M) and expanded to 350+ global companies with ARR doubling yearly; 60% of users now from marketing/e-commerce/sales departments rather than legal, indicating maturation of organizational adoption patterns. Red Points confirmed 1,300+ brand clients with 96% enforcement success rates and 100M+ daily incident processing; announced product enhancements including Brand Exposure Report with Saturation Rate metric for quantifying counterfeit risk across 86% of seller clusters operating on multiple platforms (Amazon, eBay, Instagram, TikTok Shop). RightHub's acquisition by Anaqua signaled market consolidation toward enterprise IP systems integration. However, critical evidence of persistent limitations emerged: peer-reviewed research documented algorithmic bias and jurisdictional disparities in AI enforcement systems, while practitioner assessments highlighted false positive failures eroding business trust and cooperation. Counterfeiting remained at $2.5% of global trade with sophisticated evasion tactics constraining enforcement velocity. The practice remained at leading-edge tier: production-scale deployments and multi-vendor ecosystem maturity confirmed, but structural adoption barriers (jurisdictional fragmentation, platform cooperation friction, tool reliability limitations, false positive management, organizational readiness in non-specialist departments) perpetuated the supply-demand plateau that had defined the practice for five years.
  • 2025-Q4: Vendor product maturity accelerated with major announcements from market leaders. Corsearch launched Zeal 2.0 (AI-native platform, 5,000+ customers, 46% faster enforcement) and announced TrademarkNow 2025 enhancements (neural phonetics, visual similarity, generative AI for sound-alike and logo detection). Clarivate launched RiskMark GA blending predictive and generative AI with 172.5M trademark records for instantaneous conflict detection. Red Points maintained leadership position: competitive analysis confirmed 1,300+ brand clients with 96% success rates; market data documented 59% growth in counterfeit fake websites (2023-2024) with projections of 70% growth in 2025, signaling escalating threat sophistication. Case study evidence confirmed customer ROI: Go City achieved 75% CPC reduction and 86% ad hijacking reduction using AI-powered brand protection. By year-end 2025, the practice had solidified at leading-edge tier with five-year confirmed longevity: vendor ecosystem matured with multi-generational AI innovation cycles, production-scale deployments serving 5,000+ enterprise customers across major brands, and measurable customer outcomes (cost reduction, takedown velocity, detection accuracy). However, structural adoption barriers remained entrenched: counterfeiting escalated with AI evasion tactics, jurisdictional fragmentation persisted, false positive management continued limiting trust, and organizational readiness gaps sustained the practice's confinement to specialist brand protection departments rather than mainstream enterprise IP operations.
  • 2026-Jan: Vendor platforms confirmed sustained production-scale deployment entering 2026. MarqVision reported 94%+ SKU-matching accuracy with 10M+ sellers tracked and 50% faster impersonation takedowns confirmed through customer testimonials; Red Points confirmed 1,300+ brand protection clients with 4.6M enforcements annually and 70M daily links processed. Corsearch unveiled CVAN, a proactive (pre-listing) IP enforcement capability achieving 90% moderation efficiency gains across 1,500+ monitored marketplaces, signaling evolution from reactive to proactive enforcement architectures. Strategic analysis documented organizational shift from reactive litigation toward proactive AI-powered trademark surveillance, confirming market maturation in enforcement methodology. The practice remained firmly at leading-edge tier with sustained six-year longevity: three competing vendor platforms serving 5,000+ brands, technical accuracy 94%+, and production-scale processing (70M+ daily items). However, structural barriers (jurisdictional fragmentation, platform cooperation friction, false positive management, organizational adoption beyond legal specialists) persisted unchanged, perpetuating the leading-edge plateau.
  • 2026-Feb: Enforcement velocity and scale accelerated with MarqVision's 11.35M IP infringements actioned in 2025 (101.6% counterfeit growth, 460% impersonation surge) and real-time domain takedown capability (99.8% success, 15-min MTTR). AI adoption in IP workflows jumped to 85% (from 57% in 2023), driven by operational embedding in trademark clearance. Market pressure intensified: 90% of surveyed companies now face AI-accelerated counterfeiting and impersonation threats. However, critical assessments documented persistent limitations—authentication technologies remain unreliable (Microsoft Research finding no foolproof AI-generated media detection), and trademark search tools continue facing fundamental AI accuracy gaps (Venable LLP)—underscoring the maintenance of the leading-edge plateau despite continued vendor innovation and market adoption growth.
  • 2026-Apr: MarqVision closed a $48M Series B (total $90M) targeting $100M ARR by mid-2027 with 350+ brand customers, and SnapDragon documented production enforcement at scale (Nidecker 63K listings removed at 96% compliance; Ergobaby 18K over four years). The USPTO cancelled 52,000 fraudulent or dormant marks using AI bad-faith detection tools, demonstrating government-scale institutional adoption. A ZeroFox assessment reinforced the persistent enforcement gap: AI can detect at scale but cannot enforce without registrar relationships and legal escalation pathways, sustaining the leading-edge plateau. Additional April signals confirmed production-scale maturity across the stack: Amazon's Omniscan system removed 15 million counterfeits in 2025 across 14 countries with coordinated law enforcement; Entrupy processed $3.7B in luxury goods authentication at 99.86% accuracy (33% YoY growth); and China's State Council reported 37,000 patent/trademark cases and 38,000 customs seizures totalling 75.75 million items, with 2026 priorities extending to e-commerce and livestream commerce enforcement.