loading...
Taktile built the compliance-first AI decisioning layer that lets banks automate underwriting, claims, and fraud — without waiting for regulators to catch up.
ENTRY ANGLES
Build an insurance claims decisioning platform for unstructured data (photos, medical records, invoices) with compliance-first architecture · Target mid-market digital lenders and neo-insurers with faster procurement cycles than large banks
VERTICALS
CAPABILITIES
Financial compliance expertise, MLOps, Regulatory knowledge, Enterprise sales
MAIK TARO WEHMEYER, CEO
“2026 is the year where AI comes to financial services. Models have now increased to such a strong level that they finally allow for agents to perform better than humans at many complex tasks.”
The problem with deploying AI in financial decisions is not that the models are not good enough. It is that banks and insurers cannot hand a loan application to a black box and walk away. Regulatory frameworks require documented decision logic, explicit audit trails, and human review for edge cases. The AI products that work in consumer applications fail compliance requirements before they touch a single customer in a regulated institution.
Maik Taro Wehmeyer and Maximilian Eber understood this before starting Taktile in 2020. Both had worked at QuantCo, a firm that does applied machine learning for regulated industries. Their insight: financial institutions do not primarily need better AI – they need infrastructure that makes existing AI auditable, governable, and compatible with the compliance frameworks they already operate within.
Taktile is an operating system for AI-driven financial decisions – a layer where models from any major AI lab plug in alongside human-defined rules, institutional data context, and explicit review gates for exceptions. One of the world’s largest insurers runs multiple decision workflows on the platform, projecting over $90 million in cost savings from claims processing alone. B2B loan underwriting runs at 95% automation for another customer. Anti-money-laundering false positives dropped 75% for a third. Goldman Sachs led the $110 million Series C, with Tiger Global, Index Ventures, and Y Combinator also participating. Total funding: $184 million across four rounds.
Financial institutions have spent longer at the AI pilot stage than almost any other industry. The standard explanation is regulatory risk: banks do not want to deploy AI in core processes without understanding failure modes. The more structural explanation is that AI products were not built with financial compliance as a first principle – compliance was a feature to retrofit, and it turned out to be very difficult to retrofit.
Taktile’s architecture is compliance-first by design. The platform does not ask the institution to trust the model; it asks the institution to define the decision logic explicitly, then uses models to execute that logic at scale with human review at pre-defined exception points. The AI does the work. The human defines the boundaries and reviews the cases that fall outside them. That is a fundamentally different implementation than deploying a general-purpose AI assistant into an underwriting workflow and hoping the outputs are defensible in court.
Goldman Sachs does not typically lead Series C rounds in fintech infrastructure as a purely financial bet – they invest in infrastructure companies they intend to use, or whose clients they expect to use. In regulated industries, institutional adoption creates competitive moats that revenue metrics alone do not capture.
The most immediate extension is geographic. Banking and insurance regulation differs meaningfully between the US, UK, EU, and Asia-Pacific, and Taktile’s current regulatory coverage is not entirely clear from public information. Any market where financial institutions are actively replacing manual underwriting with digital processes has the structural need.
The less obvious opportunity is in insurance claims specifically. The $90 million efficiency number came from claims processing, and claims involve far more unstructured data than credit underwriting – accident photos, medical records, contractor invoices, correspondence. A decisioning platform purpose-built for insurance claims, wrapping vision models, OCR, and medical coding AI into a compliance-first governance layer, would be a narrower but more defensible position than a horizontal platform.
The constraint for any new entrant: Taktile’s distribution depends on winning vendor approval at large, slow-moving financial institutions. With Goldman Sachs on the cap table and $184 million raised, they have the relationships and runway. A new entrant would need a different path – mid-market digital lenders, neo-insurers, or credit unions whose procurement cycles run faster and whose core banking systems are not forty years old.