Curvestone AI raises $4M to ship reliable agentic automation for regulated workflows
Curvestone AI closed a $4 million seed round to speed up product development, expand go-to-market, and grow its library of workflows. The company wants to be the trusted automation layer for agentic AI in regulated environments.
Founded in 2023, Curvestone builds workflow automation for legal, finance and insurance teams that live under audit. Think contract audits, insurance policy reviews, compliance checks, and document extraction-high-accuracy work that gets regulator attention.
Why product teams should care
The promise here is scale without sacrificing accuracy or auditability. Curvestone's no-code, building-block approach lets operations teams stand up and adapt workflows without waiting on engineering sprints, while keeping a clean audit trail.
It's LLM-agnostic and plugs into existing systems-CRM, document management, loan origination-so you don't have to force new habits on busy teams. The platform is built to keep quality consistent across multi-step workflows, reducing compound error risk that shows up in long AI pipelines.
What the platform does
- No-code "Workflows" to assemble and customize processes
- Use cases: contract risk detection, automated redlining, document comparison, data extraction, interactive Q&A
- Audit-ready by design: every review leaves a transparent, defensible trail
- Consistency controls to maintain accuracy across multi-step flows
- Works with all major LLMs and integrates with CRM/DMS/LOS tools
- Ops-led configuration so processes can adapt as regulations change
"In regulated industries, quality and scale have always been at odds," said co-founder and Chief Executive Dawid Kotur. "You can review everything and go broke, or cut corners and hope for the best. AI that actually works changes that equation by handling routine validation at scale while humans focus on the complex cases that need expert judgment."
Customers and traction
Curvestone works with law firms, mortgage networks and wealth management firms, including Stephenson Harwood LLP, Browne Jacobson LLP, Walker Morris LLP and Pivotal Growth Partners LLC.
"Curvestone is solving the hard technical problem of automating complex workflows while achieving high accuracy, and accuracy is paramount in regulated industries like financial services," said Kevin McLoughlin, partner at MTech Capital, who is joining the board. "The early traction they're seeing validates real market demand."
Funding details
The seed round was led by MTech Capital, with participation from Boost Capital Partners, D2 Fund and Portfolio Ventures.
If you're building similar capabilities
- Define accuracy targets by workflow step, not just aggregate (precision/recall, calibration, time-to-review).
- Enforce human-in-the-loop for edge cases; route by model confidence and policy rules.
- Version everything: prompts, models, datasets, and decision policies with rollback.
- Stand up an evaluation harness for regression testing across documents, models, and jurisdictions.
- Wire into existing systems of record first; avoid parallel shadow workflows.
- Log for audit from day one: inputs, outputs, prompts, decisions, overrides, and reviewer identity.
- Plan for model swaps and fallbacks; keep vendor neutrality where possible.
- Map data controls to policy: PII handling, retention, redaction, and regional routing.
Useful reference: the NIST AI Risk Management Framework is a solid baseline for controls and audits. For hands-on automation resources, see our collection on AI-driven workflow automation.
Your membership also unlocks:
