AI Governance in Asset Management: 5 Takeaways for Allocators and Managers

2. AI governance is becoming a factor in manager selection

The allocator panel made clear that AI governance, while not yet a dealbreaker in manager evaluation, is quickly becoming a differentiator. Both Nicholas Zizzo, who oversees due diligence at SEI across $1.9 trillion in assets, and Douglas Walouke, who leads operational due diligence at FEG with over 30 years of experience, described a shift in how their teams approach manager assessments.

Allocators increasingly want to understand not just whether a manager uses AI, but whether the firm has a governance structure around it, including policies, oversight mechanisms, incident response protocols, and clear documentation of where AI feeds into investment or operational decisions.

For managers, this means having a coherent AI governance story is becoming as relevant to winning and retaining mandates as cybersecurity posture became five years ago.

3. Shadow AI is the risk most firms are underestimating

John Verry of CBIZ highlighted an AI risk management problem that affects both sides of the allocator-manager relationship. According to the 2025 High Alpha SaaS Benchmarks Report, 64% of SaaS companies now embed AI as a supporting feature, and 92% have launched or plan to launch AI features. Asset management firms with fewer than 100 employees typically run 40 to 80 SaaS applications; larger firms use 150 to 300.

That means AI is already embedded in tools managers use every day, often without explicit awareness or governance. Blocking ChatGPT or Claude creates, as Verry put it, an illusion of control. The real exposure sits inside CRM platforms, compliance tools, portfolio management systems, and communication apps that have quietly added AI capabilities.

For allocators performing operational due diligence, this reinforces the need for structured, specific questions about a manager's full AI footprint, not just the tools they consciously adopted. 

4. Practical AI adoption is further along than many assume, but unevenly distributed

Herb Chain of CBIZ mapped AI adoption across four investment manager functions: research, due diligence, recordkeeping and administration, and compliance. The real-world examples he cited, including BlackRock's Aladdin AI agents orchestrating multi-step portfolio tasks, Vanguard using NLP/LLMs to predict dividend cuts from earnings call transcripts, and State Street's Alpha agents automating data checks, illustrate how quickly AI usage is maturing at the largest firms.

Mike Greenman, a Partner at Leonard Green & Partners and a member of the firm's AI Committee, offered a complementary view from the private equity side. Across 60 to 70 portfolio companies, he described AI outcomes that ranged from unexpected wins in unlikely sectors to outright failures in areas the firm had been confident AI would help. For allocators, that variability reinforces why AI-related questions in due diligence need to go beyond yes-or-no adoption checkboxes.

5. An acceptable use policy is not an AI governance program

Verry drew a clear line between firms that have published an AI acceptable use policy and those that have built an actual governance program. He recommended aligning with established frameworks like NIST AI RMF and ISO 42001, and outlined a phased 12-month approach that moves from inventory and classification, through operationalization of controls and incident response, to mature monitoring, board-level reporting, and third-party attestation readiness.

He also flagged the emerging regulatory landscape, including the EU AI Act, the Colorado AI Act, and NYC's Bias Act, which increasingly place accountability on the organization deploying AI rather than the vendor that built it.  For managers, that means AI governance carries the same liability exposure as any other operational risk. For allocators, it means the questions they ask about AI oversight need to probe for substance beyond a one-page policy document.

What this means for the relationship between allocators and managers

The consistent thread across every segment of this webinar was that AI adoption in asset management is no longer a question of if. The competitive question is whether firms can govern it well enough to maintain the trust of their investors, regulators, and counterparties.

That is why Dasseti co-hosted this event with CBIZ. As the tools and systems that managers use grow more autonomous, the due diligence and monitoring processes that allocators rely on need to evolve alongside them. Structured, scalable data collection and analysis across manager relationships is how the industry will keep pace.

Watch the webinar on demand >>