Directory of AI model risk management and AI governance vendors

AlphaBitCore sells an AI control plane that lets regulated firms govern their AI models, agents, tools and workflows in production. It enforces policy at runtime, routes actions to role based approval queues, and keeps sealed execution records that compliance and audit teams can replay to see exactly what a workflow did. For investment and wealth firms it adds a workbench of 23 agents, 95 workflows, 78 skills and 12 models for research, portfolio, advisory and compliance work, mapped to FINRA's 2026 oversight report and to the SEC's books and records rule. It deploys in a customer's own VPC or on premise, and its Nexus AI Gateway is free and open source. Paid tiers are sold by annual subscription, license or usage credits. Design partners include a credit ratings agency, a financial data provider, a wealth platform and a turnkey asset management program, none of them named. The founders come from Morningstar, IBM Research, RBC Capital Markets and TD Securities.

Block Convey builds two AI governance layers for banks, asset managers, insurers and fintechs. PRISM captures every language model call, tool invocation and reasoning step in production, scores quality, applies guardrails that strip regulated identifiers at ingestion, and exports regulator-facing evidence packs in under a minute from tamper-proof immutable traces, with compliance teams reviewing session-level transcripts rather than raw logs. PRISMX is a managed browser extension enforcing data loss prevention on consumer assistants, intercepting paste-and-prompt traffic at the endpoint so personal data, health data and source code never reach an external model. Obligations are mapped individually across US banking model risk guidance, state cybersecurity rules, fair lending requirements, insurance model bulletins and European AI and operational resilience regimes. Supported models, cloud platforms and agent frameworks are named explicitly.

Ethos builds an end-to-end platform for model risk management at banks and fintechs, covering the full supervisory lifecycle from model development and documentation through validation, reporting and governance. It gives institutions real-time visibility and automation across their whole model inventory, and is designed to handle both conventional statistical models and newer machine learning and generative systems, which is the gap most existing frameworks have: institutions have deployed machine learning faster than their model risk functions could absorb it, and model governance is now a primary focus for United States banking examiners. The company positions itself against the decisions models actually drive at financial institutions, spanning lending, loss forecasting, fraud detection and anti money laundering. Founded in 2023 and backed by two financial services specialist funds alongside a major bank's venture arm.

EVE AI Core is a deterministic control plane that intercepts every AI action before execution rather than reviewing it afterwards, returning an allow, block or modify verdict against versioned policy packs in under a millisecond, fail-closed and with no language model anywhere in the decision path. Its argument is that a safety layer built on a model can itself hallucinate, so a non-deterministic safety check is not a safety check. Every verdict emits a cryptographically signed certificate bound to the exact policy version in force, appended to hash-chained trails, replayable on demand and verifiable offline by a third party without the vendor in the loop. It governs agent actions rather than only prompts and responses, refusing unregistered tools, risk-scoring each step and requiring human approval for high-stakes actions. It positions itself as the runtime enforcement and evidence layer beneath an existing model risk program, for lending, insurance and trading decisions.

Fiddler AI is a Palo Alto based AI observability and agent control plane vendor that sells a separately addressed financial services line rather than a financial services page. The platform evaluates, monitors, enforces and governs predictive models, generative applications and agents, whether first party, third party or coding agents. It works through the gateway an enterprise already runs, capturing every prompt, tool call and outcome, and produces unified audit trails and compliance reporting across the agent lifecycle. The financial services capability is specific rather than adapted. Credit, lending and underwriting model monitoring comes with drift root cause analysis and a slice and explain function for segment level feature impact. Automated lending decisions are explained using Shapley values and a proprietary variant, at both local and global level, with what if analysis on prediction outcomes. Fraud detection monitoring is tuned for heavily imbalanced datasets and alerts on anomalies in real time. Credit card and payment default risk ships with fairness metrics out of the box, namely disparate impact, group benefit, equal opportunity and demographic parity. Robo advisory drift detection runs against market volatility and asset class performance. The agentic line covers multi agent lending systems that assess collateral and recommend terms, collections agents that negotiate payment plans, and financial crime investigation agents. Real time moderation blocks agent conversations for policy breaches and detects personal data leakage before exposure, and human approval is enforced on high value loans. Evaluation runs on the company's own Trust and Centor model families, which execute inside the customer's virtual private cloud, so governing generative agents requires no data sharing and no external model interface calls. Published work includes a Fortune 100 financial services institution spanning wealth management, brokerage and asset management, and a named data science leader quoted on production monitoring.

ibl.ai sells an AI agent platform that institutions run themselves, in their own cloud, on premise or fully air gapped, with the full code licensed perpetually to the customer. For banks, broker dealers, asset managers and wealth firms it offers agents for compliance monitoring, risk analysis, KYC and anti money laundering checks, fraud detection, client advisory, portfolio analysis, onboarding and trading operations. The platform works with Claude, GPT, Gemini, Copilot, Amazon Bedrock, NVIDIA and open source models, chosen and switched by the customer. It connects to core systems such as FIS, Fiserv, SS&C, Jack Henry and Broadridge, CRMs such as Salesforce Financial Cloud, Redtail and Wealthbox, and data from Bloomberg, Refinitiv and FactSet. It stresses documented models, version control, independent validation and audit trails for model risk reviews, and says it is SOC 2 Type II compliant. Enterprise pilots start at $15,000. ibl.ai serves more than 400 organizations across education, government, healthcare and finance, though it names no financial institution. It is family owned and based in New York.

Monitaur is a Boston based AI governance software company selling to highly regulated enterprises, with its deepest expertise in insurance and a growing financial services segment. The platform is organized on a policy to proof roadmap. It runs from policy definition and taxonomy through model inventory, stakeholder collaboration and lifecycle governance execution, to continuous monitoring, validation and audit ready reporting, so that a governance framework becomes checkable evidence rather than a document. The distinguishing asset is a validated controls library that a carrier maps to the policies and frameworks it already operates against. It ships with alignment to named regulatory instruments rather than generic principles: the NAIC model bulletin, the NIST risk management framework and the EU AI Act. Alignment also covers model filings, market conduct examination, own risk and solvency assessment, actuarial standards of practice and the individual state insurance departments. Capabilities include model validation, continuous fairness and bias stress testing, performance and quality monitoring across multiple modeling paradigms, and governance of third party and vendor supplied models. An enterprise insurer customer cited that last capability as the differentiator because it integrated with existing third party risk management workflows. Named customer work includes the property analytics firm CAPE Analytics, whose general counsel is quoted on the selection, and case studies covering an enterprise insurer and a Fortune 200 financial services and insurance company. The advisory board includes a former director of the Arizona insurance regulatory agency who co chaired the innovation and technology committee at the NAIC. The company also has an alliance with a large professional services firm in Germany covering European regulatory compliance.

Prism Layer encodes an organization's risk frameworks, regulatory context and institutional knowledge into what it calls a governed agentic execution layer, aimed at financial services, payments and fintech firms. Its distinguishing claim is that rather than storing risk records after decisions are made, it executes the structured reasoning behind risk analysis as those decisions happen, producing auditable and defensible output at business speed. Supported work covers risk assessments, key risk indicators, controls testing, risk and control self-assessments and product risk review. The company emerged from stealth in April 2026 with a pre-seed round led by a venture firm focused on financial services and regulatory technology whose partners include former senior leaders of the US consumer financial regulator. Its three co-founders were all risk executives at the same payments company, with prior enterprise risk leadership across money transfer, brokerage, card network and consumer lending businesses, and direct experience with Irish, Australian and UK regulators.

ValidMind automates model documentation, testing and validation for bank model risk management teams. Independent comparisons describe it as the only platform built exclusively for financial institution use rather than adapted from enterprise AI governance. It maps directly to the supervisory regimes that govern this work across four jurisdictions, covering United States interagency guidance old and new, the United Kingdom prudential regulator's principles, the Canadian supervisor's model lifecycle requirements and the EU AI Act, and produces audit ready evidence against each. Recent work extends the same framework to autonomous agents, recording who approved an agent, under what conditions and at what risk tier, with real time policy enforcement rather than periodic review. A major credit bureau has embedded the platform inside its own analytics environment so banks can document credit and fraud models against several regimes at once.