A purpose-built, Zero Data Retention AI workspace delivered for Insigneo, an SEC and FINRA-regulated wealth management firm, giving its executive leadership secure, governed access to generative AI.

Project Overview

Insigneo, an SEC and FINRA-regulated broker-dealer and wealth management firm, wanted its executive team to benefit from generative AI without compromising its regulatory obligations. Off-the-shelf tools, including Claude Enterprise’s native chat interface and Microsoft 365 connectors, could not meet the firm’s data retention requirements creating a clear gap between AI adoption and compliance.

Wildnet Technologies designed and delivered the Insigneo Executive AI Assistant: a purpose-built, Insigneo-branded AI workspace for its executive leadership team, powered by Anthropic’s Claude API with Zero Data Retention enforced end-to-end.

Business Challenge

Regulatory obligations under SEC and FINRA rules meant client and regulated data (PII, NPI, KYC/AML) could not be retained by a third-party AI vendor
Claude Enterprise’s chat interface and native Microsoft 365 connectors are not ZDR-eligible, ruling out an off-the-shelf deployment
Executives needed a simple, personalized experience comparable to a consumer AI assistant, without compromising compliance
No existing governance framework existed for how leadership should responsibly use generative AI

Solution

Wildnet built a custom AI workspace that separates two concerns: stateless AI inference through the Claude API under a ZDR arrangement (so nothing is retained on Anthropic’s infrastructure), and memory and continuity managed entirely within Insigneo-controlled Azure infrastructure. Key components:

ZDR architecture: all requests routed through the Claude Messages API under a Zero Data Retention arrangement, ensuring no Anthropic-side retention of prompts or outputs
Protected Memory layer: Azure PostgreSQL with pgvector, Azure Functions, Azure SQL for RBAC and audit, and Azure Key Vault for encryption keeping context within Insigneo’s own environment
Prompt Classification Engine: detects PII, NPI, KYC/AML data, and financial identifiers (CUSIP, ISIN), including scanning of uploaded PDF and Word files
Governance: role-based access control, full audit logging, and alignment with SEC, FINRA, and NIST AI RMF frameworks

Delivered in four phases — Foundation, Build, Training and Rollout, and ongoing Governance — with dedicated AI governance, architecture, and adoption roles working alongside Insigneo’s CTO, IT, and compliance stakeholders.

Technology Stack Used

Claude API (Messages API) with Zero Data Retention
Microsoft Azure hosting
Azure PostgreSQL with pgvector
Azure Functions
Azure SQL
Azure Key Vault
custom Prompt Classification Engine
custom Microsoft 365 integrations
SSO/SCIM identity integration

Claude Products Used

Claude API (production, ZDR-enabled)
Claude Code

Measurable outcomes

61 executive users including tester onboarded to a compliant, ZDR-enforced AI workspace
Zero compliance complaints or incidents since go-live
Sustained growth in adoption and usage
10.6M tokens processed in the most recent 7-day period, reflecting active, ongoing executive use across Management, Finance, and Compliance functions

Client Review

“Bringing generative AI into a regulated environment felt risky at first, but the compliance-first approach changed that. Every step, from data protection to access controls, was built with our regulatory obligations in mind. Our leadership team now has secure, governed access to AI that fits how each function actually works, and the ongoing governance support means we stay audit-ready without extra overhead. It’s been a smooth, low-risk way to bring our executives into the AI era.”
— Vikas Saxena, Chief Technology Officer, Insigneo

Nitin Agarwal

Managing Director (MD) Nitin Agarwal is a veteran in custom software development. He is fascinated by how software can turn ideas into real-world solutions. With extensive experience designing scalable and efficient systems, he focuses on creating software that delivers tangible results. Nitin enjoys exploring emerging technologies, taking on challenging projects, and mentoring teams to bring ideas to life. He believes that good software is not just about code; it’s about understanding problems and creating value for users. For him, great software combines thoughtful design, clever engineering, and a clear understanding of the problems it’s meant to solve.

Insigneo Executive AI Assistant: Compliant Generative AI for a Regulated Wealth Management Firm

Table Of Content