Flank

Flank is a Berlin-based agentic AI platform for enterprise legal teams that automates repeat legal, compliance, and security work inside the tools business users already use. Public product materials position it as ‘frontline legal agents’ that triage requests, draft and negotiate NDAs and other routine agreements, answer legal and compliance questions, and complete questionnaires through email, Outlook, Teams, Slack, browser workflows, Jira, Salesforce, and CLMs. The best current market proof comes from June 2025 funding coverage and legal-industry press: Flank raised a $10M round led by Insight Partners with participation from Gradient Ventures, HV Capital, and 10x Founders, bringing total funding to roughly $18M; Tech Funding News says customers include DeepL, SumUp, TravelPerk, QA, PROS, Lusha, and Simmons & Simmons, and quotes TravelPerk’s CLO saying the system handles 5,000 requests per month. The Law Society Gazette separately reports that Simmons & Simmons is introducing agentic AI into its legal and compliance teams using Flank alongside the firm’s Percy model, with NDAs, forms, and policy/compliance queries as concrete examples. Vendor-claimed security posture is strong on paper: SOC 2 Type II, regional tenancy, zero data retention, no model training on customer data, SSO with major identity providers, audit logs, and approval thresholds. The main caveats are evidence quality and fit. Public pricing is absent, most product and usage detail still comes from Flank’s own materials or interview-based press, and Reddit/review-site signal is effectively nonexistent because the brand term is heavily polluted by non-legal search intent. This looks like a serious in-house legal-ops product, but still one that buyers would need to diligence directly rather than trust on community signal alone.

Company Info

• Founded: 2018
• Funding: $18.2M
• HQ: United States
• Sector: Compliance, Cybersecurity

What We Haven’t Verified

This page was assembled from publicly available information. Feature claims and workflow mappings are based on what the vendor and third-party listings publish — not hands-on testing or practitioner feedback.