Key Intelligence Insight
Taktile is not a rules engine. It is not an AI vendor. It is decisioning infrastructure for the financial institution, and that distinction is the entire thesis.
Banks and fintechs have spent decades making critical automated decisions by encoding logic in backend code or buying heavy legacy platforms that domain experts cannot touch. The result is a translation layer between the people who understand risk and the systems that execute it. That layer is full of delays, miscommunication, and slow iteration. Taktile eliminates it.
The product launched in its current form in January 2023. In two years it quadrupled its customer base, tripled revenue, and closed Nubank, Mercury, and Allianz. In February 2025, it raised a $54M Series B in ten days with six competing termsheets. The speed of that raise signals something the market rarely produces: a category that is clearly being won.
The structural bet is consolidation. Financial institutions currently buy point solutions for KYC, credit, fraud, AML, and collections. These point solutions do not talk to each other. Taktile's thesis is that all of these decisions share the same underlying architecture: data in, rules and models applied, decision out. One horizontal platform captures all of it. That is the moat being built.
Founding Story
Maik Taro Wehmeyer and Dr. Maximilian Eber are both machine learning engineers. Before Taktile, they built a consultancy modeled loosely on Palantir: recruiting strong quantitative talent and deploying them against hard modeling problems at large enterprises. Insurance companies, banks, e-commerce. Every industry, every type of ML problem.
That work taught them two things. First, the decisions that matter most in financial services are overwhelmingly automated. Second, the infrastructure supporting those decisions is decades behind the models being built on top of it.
They entered Y Combinator in 2020 with a different thesis: an ML deployment platform for every industry. The cloud providers killed it. AWS, Google, and Databricks built directly competitive products. Winning against that surface area, at that price point, was not a fight worth having.
The pivot was not a reinvention. It was a sharpening. Wehmeyer and Eber had always believed in mission-critical decisioning as the target. They narrowed the product to financial services, expanded the architecture from pure ML deployment to a full decisioning layer including rules, case management, data integrations, and optimization tooling, and relaunched in January 2023. They threw away 200,000 lines of code to do it.
Product-market fit came quickly. The Series B followed two years later. Balderton Capital led. Index Ventures and Tiger Global participated.
Product
Taktile is a SaaS decisioning platform. Financial institutions use it to build, test, monitor, and optimize automated decisions across the customer lifecycle: onboarding and KYC, credit underwriting, fraud detection, AML and compliance, claims management. The platform sits in the background of every critical risk event.
The architecture has two layers.
Agentic Platform Core
The Decision Engine is the foundation: a low-code visual builder for decision logic that non-technical teams can operate directly. Risk analysts, credit teams, and fraud managers configure and iterate without engineering tickets. An AI Copilot writes and debugs logic. Backtesting and A/B testing tools let teams run counterfactuals on historical traffic before pushing changes to production. That optimization loop is rare. Most in-house systems and legacy platforms never close it.
The AI Agent Manager deploys pre-built and custom AI agents tuned to financial services use cases: onboarding completeness checks, adverse media investigations, sanctions match review, financial spreading. Each agent operates inside a governance framework with human-in-the-loop controls and full audit trails. The Case Manager connects directly to agent workflows, routing only the cases that require human judgment to the right analyst.
The Context Layer is the data orchestration surface. Over 200 plug-and-play integrations connect to credit bureaus (Equifax, Experian, SCHUFA, Serasa, Transunion), identity and fraud providers (Socure, Socure, Sentilink, Veriff), open banking rails (Plaid, TrueLayer, Tink), AML tooling (ComplyAdvantage, OpenSanctions), business intelligence providers (Dun & Bradstreet, Middesk, Companies House), and more. A fintech expanding into a new geography needs multiple bureau integrations. Taktile already has them. The integration maintenance cost shifts off the customer entirely.
Domain-Specific Solutions
On top of the core platform, Taktile ships pre-configured agent workflows for specific use cases. SMB underwriting agents process up to 5x more loans by automating financial spreading, credit memo generation, and document classification. AML agents reduce false positives by automating transaction monitoring review and regulatory report generation. KYC/B agents manage application completeness checks and business ownership intelligence. Insurance underwriting and claims agents scale capacity without headcount growth.
The product strategy is clear: own the horizontal infrastructure, then layer domain-specific acceleration on top. Every new agent template is a distribution advantage. Every new data integration reduces a competitor's ability to catch up.
Market, Competition & Business Performance
Market
Financial services will spend $97 billion on AI by 2027. The gap between what general-purpose AI can do and what regulated institutions can actually deploy remains wide. Taktile's applied research arm, Taktile Labs, frames this as its core problem.
The addressable market compounds in multiple directions. The immediate TAM is banks and fintechs replacing legacy decisioning infrastructure. That is already large. Beyond banking, Taktile has closed customers in insurance, energy, and telco. Every asset-based business carries credit default risk. Every company that issues a phone or provides energy on credit is running an underwriting decision. The decision architecture is the same. The surface area expands substantially.
The structural tailwind is the migration away from in-house builds. Nubank spent twelve years building its own decisioning stack. It switched to Taktile. Mercury, one of the most technically sophisticated fintechs operating today, chose Taktile. The investor hypothesis that sophisticated players would never outsource their decisioning infrastructure has been empirically disproved. That inflection point is recent and compounding.
Competition
Taktile competes on two axes. The first is point solution vendors: fraud tools, credit underwriting engines, KYC platforms, AML monitoring software. These are Oscilar, Alloy, SEON, and a range of bureau-adjacent products. They are all good at one use case. The consolidation thesis runs directly at them.
The second axis is process automation platforms: Camunda, UiPath. These players own workflow orchestration and are pushing toward financial services applications. The overlap is real. Wehmeyer names UiPath as a structural analog and acknowledges the competitive surface area will grow. Taktile has a published integration with Camunda through the Taktile Connector on the Camunda Marketplace, suggesting a partnership-first posture rather than direct displacement.
The cloud providers remain a latent risk. They killed Taktile's first product. Their current platforms are horizontal, not domain-specific. A cloud provider building deep financial services decisioning expertise is theoretically possible. It has not happened. The domain proximity required to compete in credit, AML, and fraud is a meaningful barrier.
The competitive question is whether being best in class for risk decisioning compounds into a business that displaces the point solution category before that category consolidates through M&A.
Business Model
Taktile sells enterprise B2B SaaS on multi-year contracts. Pricing scales with the volume and complexity of decisions running through the platform. Larger institutions, longer contract terms: two to three years for mid-market, five years for regulated institutions that avoid system changes.
The go-to-market motion is consultative and pod-based. Account executives pair with solution engineers and customer value engineers against specific account lists. The solution engineering function, led by Annie Wachsmuth with a 30-person team, manages the full customer lifecycle from pre-sales demo to technical implementation to renewal. Time to go live targets three months. Contract preparation for renewal starts three to six months in advance.
The unit economics are structurally strong. Wehmeyer reports zero customer churn from companies that did not go bankrupt. Customer lifetime value is effectively uncapped. Net dollar retention compounds through expansion: customers that start with one use case add additional use cases over time, increasing ACV without incremental acquisition cost. The platform stickiness is genuine. Replacing the decisioning layer is replacing the heart of a financial institution's risk infrastructure. Nobody does it twice.
Traction
Taktile launched its current product in January 2023. By mid-2025 it had surpassed 200 financial institution customers, having quadrupled its base in 2024 alone. Revenue tripled in 2024. The Series B closed in February 2025 at $54M, led by Balderton Capital, bringing total funding to $79M.
Named customers include Nubank (the world's largest neobank), Mercury, Zilch, Allianz, Allianz Partners, Rakuten, and Pleo. This is the strategic sequence Wehmeyer describes as lighthouse logo compounding: sophisticated early adopters validate the category, large enterprises follow, ACV scales.
G2 named Taktile a Category Leader in Decision Management Platforms for Spring, Summer, and Fall 2025. The Equifax reseller partnership provides credit teams with direct Equifax data access inside the platform. The Prism Data partnership adds real-time cash flow underwriting for lenders. The MANTL integration connects Taktile's credit decisioning into community bank and credit union account opening workflows.
Taktile Labs published FinSpread-Bench in February 2026, the first public benchmark for agentic AI performance on financial document processing. GPT-5.2 with Gemini extraction reached 96.5% field match rate, above the 89% human baseline. The research function is not decorative. It is building the evaluation infrastructure that regulated institutions need before deploying AI in production. That infrastructure becomes a distribution advantage.
The company operates from offices in New York, London, Berlin, and Iasi. Headcount sits at 130 to 150 employees. The stated ambition is a NYSE listing.