Most asset management organizations are past the question of whether to adopt agentic AI. The real question sitting in front of CFOs, COOs, and CTOs right now is narrower and harder: can the organization prove the reliability of its AI-assisted work when an auditor, examiner, or investor asks how a number was produced?
What is agentic AI, and why does it matter to asset managers?
Agentic AI differs from traditional automation. While the traditional approach executes a single rule or answer a single prompt, an agent can carry out a multi-step task, such as validating NAV inputs, flagging exceptions, and drafting a variance memo, while routing judgment calls to a qualified person at defined checkpoints.
That distinction matters because asset management operations are full of multi-step, judgment-heavy processes. NAV calculation pulls from fund administrators, pricing services, and portfolio systems, then requires a human to explain what moved and why. Portfolio valuation requires normalizing inconsistent data before a committee can even review it. These aren’t single-click automations. They’re workflows that agentic systems are built to manage.
Adoption is accelerating, but most organizations haven’t crossed into production. Mercer’s 2026 AI in Asset Management Survey found 55% of asset managers have AI integrated into at least one investment process. However, 27% are still in the pilot stage, and 18% haven’t started.
The gap between experimenting and scaling is where real opportunity lies, but it’s also where governance determines whether AI initiatives succeed, stall, or create unintended risks.
Why governance is the real barrier to scale
Most AI efforts stall between the demo and production. Data is worse than promised. Controls arrive late, if at all. No one can quantify the value of the pilot. For a regulated asset manager, that’s not just an inconvenience. It’s a liability.
The pressure is concrete. An auditor wants the evidence chain behind a reported NAV or a hard-to-value mark. A limited partner presses on how a fee was allocated. An examiner asks for documented controls over any AI agent touching a material process. The SEC has already brought several “AI washing” cases against advisers who marketed AI capabilities they didn’t have, resulting in enforcement actions and billions of dollars in penalties since 2021.
Organizations that plan for governance from the start see a different pattern. Portfolio risk and value become visible in real time. Audit readiness comes from a documented AI inventory instead of a scramble. Governance, done well, is not a brake on adoption. It’s what makes adoption defensible enough to scale.
Build AI Agents with Governance
In most organizations, adoption has consistently outpaced oversight. The better approach builds controls into the first agent rather than bolting them on after go-live. Every cluster keeps a qualified person in the loop at defined checkpoints, because a NAV, a valuation, or a filing still must be defended by a person when the auditor, examiner, or investor asks.
This approach follows the National Institute of Standards and Technology (NIST) AI Risk Management Framework and the governance guidance from the Committee of Sponsoring Organizations of the Treadway Commission (COSO). For a regulated manager, that’s the difference between a demo and a capability that holds up in an SEC exam, an external audit, and investor diligence.
How an AI Jumpstart proves value
Rather than launching a multi-month program, a Jumpstart serves as a proof of value to deploy a single agent on an organization’s own data and systems and produces a concrete deliverable in the first few weeks. That might be a data quality scorecard, a NAV variance memo explaining what moved and why, or a valuation lineage report tracing every mark back to its source.
You see real output before committing to anything larger. From there, the agent expands into a full cluster, sequenced by an orchestrator that decides what runs and in what order. A cleanup that used to happen once becomes a standing process. In portfolio valuation, for example, normalization that consumes days of manual work becomes a step that runs on demand, with a documented trail behind every number.
The AI Field Guide for Asset Managers
To help asset managers build AI into their operations, CrossCountry created the AI Field Guide for Asset Managers. This resource provides a map of where operations break down across the fund lifecycle and how a risk-tiered governance model turns AI adoption into a defensible, repeatable advantage, rather than a compliance afterthought.
The AI Field Guide maps 12 agentic clusters across three areas of the operating model:
- Revenue, valuation, and data integrity: NAV calculation, fee and expense allocation, portfolio valuation, and the data foundation underneath all of it.
- Reporting, performance, and liquidity: the close, investor and board reporting, portfolio monitoring, and capital planning.
- Governance, compliance, and treasury: controls monitoring, investor and legal obligations, regulatory operations, and treasury.
Each cluster is built for how asset managers operate, including legal entity hierarchies, investor agreements with bespoke terms, and the audit and regulatory rules layered onto every process. The fixes are specific because the operational breaks are specific.
To understand how a Jumpstart fits within a broader AI Strategy & Transformation roadmap, the field guide maps the full sequence from pilot to production.
Get started
Most organizations begin with a short discovery session, select one or two clusters tied to the outcome that matters most, and stand up a Jumpstart in weeks. The agents are aimed at the numbers that decide a fund’s standing: NAV accuracy, the returns investors underwrite, operating margin, regulatory readiness, and the length of the close.
The question is no longer whether agentic AI belongs in asset management operations. It’s which break to fix first, and whether the output can hold up when it’s challenged.
For asset managers, the next step is not another pilot. To launch your own Jumpstart, contact CrossCountry today.