Research Engineer - Evals

Variance, building AI investigative agents for risk and compliance, today announced a $21.5 million Series A funding round to expand its agentic AI platform ...

Funding, investment, acquisitions & ipos - Q2 2026. Q2 2026 has wrapped up! Time for its quarterly investment update. Explore the latest funding rounds and acquisitions in the world of fraud prevention, identity verification, and risk mitigation - a quarter defined by a landmark raise in AI-powered decision automation, a wave of Series A capital flowing into compliance tooling across Europe and Asia, and continued platform consolidation through strategic acquisitions. Funding. Taktile. Berlin- and New York-based Taktile raised $110 million in a Series C round led by Goldman Sachs Alternatives (Growth Equity), with participation from existing investors Balderton Capital, Index Ventures, Tiger Global, and Y Combinator. The company builds an AI-powered decision automation platform purpose-built for financial institutions, enabling them to automate high-stakes decisions across loan underwriting, insurance claims processing, customer onboarding, fraud detection, KYC, and AML compliance - all within a governed, auditable framework designed to meet regulatory scrutiny. Taktile's customers have reported outcomes including 95% automation in B2B underwriting and a 75% reduction in AML false positives. The Series C marks one of the largest raises of the quarter in the fraud and compliance technology space, reflecting strong investor conviction that the financial services industry is ready to move beyond rules-based decisioning and into AI-native orchestration at scale. Cinder. Cinder raised $41 million in a Series B round to expand its AI trust and safety platform, which is designed to detect and mitigate abuse across digital platforms and financial services. The company helps organisations identify harmful patterns - including fraud rings, coordinated inauthentic behaviour, and policy abuse - using machine learning models that operate at the speed and scale that manual review cannot match. The round reflects growing recognition among investors that trust and safety infrastructure is becoming a critical layer of the modern digital economy, particularly as AI dramatically lowers the cost of launching sophisticated abuse campaigns. Variance. Variance raised $21 million in a Series A round led by Ten Eleven Ventures, with participation from 645 Ventures, Y Combinator, Urban Innovation Fund, and Okta. The company builds AI agents for financial crime investigations, focusing specifically on identity investigations, AML compliance, and enhanced due diligence. Variance's agents are designed to be 10 times more accurate than conventional models and 100 times faster than manual review, while maintaining the auditability regulators require. Customers have reported 50% more consistent enforcement of policies versus manual investigations. The raise reinforces investor appetite for agentic AI that is purpose-built for regulated industries - not adapted from general-purpose models, but engineered from the ground up to be explainable, auditable, and compliant. Prelude. French startup Prelude raised $20 million in a Series A led by 20VC, with participation from Singular, Seedcamp, Deel, FDJ UNITED Ventures, and a group of angel investors including Steffen Tjerrild, Antoine Le Nel, and George Arison. The company is building AI-powered trust signals to replace legacy verification mechanisms like CAPTCHAs and SMS one-time passwords - authentication methods that are increasingly vulnerable to bot attacks, SIM swapping, and AI-generated bypasses. Prelude's technology distinguishes human from automated traffic at the session level, providing financial institutions and digital platforms with a more robust first line of defence against bot-driven fraud. The round signals growing appetite for next-generation identity assurance infrastructure as the limitations of traditional friction-based security become more pronounced. Mokn. Paris-based identity protection startup Mokn raised $15 million in what marks a notable milestone: Google Ventures' first-ever investment in a French startup. The company is building technology to detect and prevent identity theft before it happens - intervening earlier in the attack chain than traditional identity verification and fraud detection tools, which typically act after credentials have already been compromised or misused. The GV backing underscores both the strategic importance of early-stage identity protection and growing transatlantic interest in Europe's fraud prevention ecosystem, particularly as French and EU regulators push for higher digital identity standards under frameworks like eIDAS 2.0. Flagright. Singapore-based Flagright raised $12.5 million in a Series A led by Infinity Ventures, with participation from Sella, Frontline Ventures, and Y Combinator. The company provides an AI-powered compliance platform that handles the full compliance lifecycle for financial institutions - transaction monitoring, screening, risk scoring, investigations, quality assurance, and regulatory filing - with explainable AI at its core so every decision can be traced and justified to regulators or auditors. Customers report payback in 4.6 months, up to 93% fewer false positives, and up to 81% lower compliance operating costs, with deployment timelines as short as two weeks. The raise reflects continued investor confidence in modular, API-first compliance infrastructure that financial institutions can deploy rapidly without displacing legacy systems. Guzco. Dutch startup Guzco raised €1 million in a pre-seed round backed by Arches Capital to automate disputed payment claims for e-commerce merchants. The platform addresses what the company describes as a $132 billion annual risk to global e-commerce, tackling friendly fraud, package non-delivery claims, chargebacks, BNPL disputes, and consumer claims of non-receipt through AI-driven automation. Guzco already works with payment processors including Mollie, MultiSafepay, and PostNL, and is expanding across Europe. The raise signals early-stage investor interest in dispute automation as a distinct category within the fraud stack, particularly as BNPL penetration and cross-border e-commerce volumes continue to grow. Fraudio. Portuguese fraud prevention firm Fraudio closed a new funding round - amount undisclosed - led by Alea Capital Partners, with participation from IMGA Private Equity. Founded by João Moura, Fraudio provides machine learning and behavioural analytics to detect suspicious activity in milliseconds for financial institutions, fintechs, payment service providers, digital banks, and payment platforms. New clients include GnosisPay and PayTabs Global, and the company is planning a 12-18 month expansion into the Americas, APAC, and MENA regions. Alongside the raise, Fraudio announced plans to expand detection capabilities for mule account networks and fast-spend behavioural attacks - two of the fastest-growing fraud vectors targeting European payment infrastructure. Acquisitions. Incode / Identiq. Identity verification platform Incode has acquired Identiq, a specialist in privacy-preserving cryptographic solutions for peer-to-peer anti-fraud collaboration, as part of a broader $100 million commitment to advancing privacy-preserving identity infrastructure. Identiq's technology enables organisations to share fraud signals with one another without exposing underlying customer data to third parties - eliminating the need for central data lakes or data brokerage intermediaries that introduce both privacy risk and regulatory exposure. The company spent nearly a decade and over $50 million developing its cryptographic approach to consortium-based fraud intelligence. For Incode, the acquisition deepens its capability beyond point-in-time identity verification toward continuous, collaborative trust allowing its platform to draw on network-level fraud intelligence while preserving the privacy guarantees that regulated businesses increasingly require. The deal reflects a broader industry conviction that the future of fraud prevention is not siloed detection, but privacy-safe collaboration at scale.