The demand on financial institutions to do more with fewer people is a familiar refrain — but the way firms are answering it is changing fast. Instead of bolting on generic copilots, risk and operations teams are reaching for AI that can actually make and defend decisions.

That means automating high-volume calls, trimming false positives and delivering quicker, fairer customer outcomes without hiring more staff. The shift is from generic assistants toward agentic systems that can parse documents, read policies, run checks and produce auditable decisions on the spot — and it is spawning a whole new category of infrastructure to support it.

A new layer beneath the decisions

The platforms emerging now let credit, fraud, risk and operations teams build, test and run AI-driven workflows without leaving the compliance guardrails a regulated business demands.

To work in this setting, these tools have to fuse frontier models with business rules, data connectors and human oversight, so every outcome — whether a person or an agent made it — stays compliant, explainable and tied to business objectives. Taktile is staking out a central spot in that stack. Founded across Berlin and New York, the company sells an Agentic Decision Platform built to automate complex, high-stakes workflows spanning underwriting, fraud and claims. By coordinating several specialised AI agents alongside rules and data, it lets teams move from manual review to automated, audit-ready decisions without a heavy engineering build.

What the platform actually does

At its core, the system blends AI agents, business rules, data connectors and human checkpoints so institutions can automate intricate decisions that still hold up to an audit.

Its use cases stretch across SMB and B2B underwriting, KYC and AML operations, transaction monitoring, identity verification and insurance claims. Because it orchestrates multiple purpose-built agents, the platform can ingest documents, interpret policy coverage and reach a payout decision far quicker than a manual team. The company's chief executive frames the leap plainly: given the right context, AI agents can now decide whether to lend to a business or onboard a customer with a speed and accuracy that would have been out of reach a year earlier. The harder task, he argues, is making each of those automated decisions controlled, auditable and fit for the most heavily regulated industry there is.

General-purpose AI is fine for simple automation, but it falls short of running mission-critical financial decisions where a single error can cost millions. — Taktile's chief executive

Money, momentum and marquee customers

The commercial story behind the technology has accelerated sharply over the past year.

In June 2026 the company closed a US$110m Series C led by the growth-equity arm of Goldman Sachs Alternatives, with a roster of established venture backers joining in. That lifts total funding to US$184m, earmarked for product development, global expansion and deeper vertical tooling for banking and insurance. The customer list carries weight too: business-banking and fintech names such as Mercury, Monzo, Faire and Pleo are on board, and the company says one of the world's largest insurers is running several use cases with projected savings topping US$90m in claims processing alone. Reported results include 95% automation in B2B underwriting and a 75% reduction in AML false positives — evidence, the firm argues, that agentic workflows can cut cost and sharpen risk outcomes at the same time. Thousands of employees still process these decisions by hand today; the pitch is to redeploy that capacity to higher-value work while keeping every outcome optimal for both business and customer.

An ecosystem of data and fraud signals

Raw decisioning is only as good as the data feeding it, so the company has assembled a network of reseller and integration partners to plug best-in-class signals straight into workflows.

A US reseller arrangement with Equifax lets customers tap credit-bureau data faster for richer cash-flow and alternative-data underwriting. A SentiLink partnership gives banks and fintechs immediate access to synthetic-identity and first-party fraud detection without custom builds, while an Inscribe integration allows agentic checks on IDs, bank statements and supporting documents inside the platform. Mifundo's verified cross-border credit data helps European lenders make faster, consistent and compliant decisions for borrowers with international histories, and an OpenSanctions tie-up strengthens AML screening with real-time access to more than two million sanctioned and watchlisted entities.

Planting a flag in Latin America

Alongside the raise, the company is putting real money behind regional growth.

It has allocated US$20m to speed AI transformation for financial institutions across Latin America, including a new São Paulo office and regional leadership hires. The move tracks rising demand from banks and insurers in the region for automated underwriting, fraud prevention and claims workflows that can scale without a matching jump in headcount — the same pressure driving adoption in more mature markets, now surfacing in a fast-growing one.