In May 2026, Fitch Ratings surveyed global CLO managers on artificial intelligence and found broad agreement that it now matters strategically. What managers are actually doing with it is narrower than the headlines suggest. It is a picture of AI in investment management as support work, not decision making.
According to Fitch’s report, the deployment is concentrated in research support, document processing, surveillance, and workflow automation. Nobody is handing credit decisions to a model.
Fitch has a phrase for the pattern: AI assisted, human decided. Machines prepare information, route work, and surface exceptions. Experienced people review the output and decide.
That is a less exciting story than autonomous trading, and a considerably more useful one, because the boring parts of running a CLO fund are where the hours actually go.
Why Document Work Is the Obvious Starting Point
A collateralized loan obligation holds a portfolio of senior secured corporate loans and finances it by issuing debt and equity tranches with different risk profiles. The manager selects and trades the loans, monitors the portfolio, and keeps the vehicle inside its governing tests. Cash flows to investors through a defined payment waterfall.
Each of those pieces generates paper. Every underlying loan has a credit agreement and an amendment history. Deal-specific compliance tests need continuous tracking. The waterfall only pays correctly if the terms and data behind it are right.
Multiply that across several deals, add trustee reports, agent notices, amendments, and internal records, and the operating burden compounds fast. Information preparation, not investment judgment, is what consumes the day.
That is precisely the shape of work AI handles well.
The Four Places It Gets Used
Fitch’s categories translate into four practical areas inside fund operations:
Document processing. A first pass over long credit agreements and amendments, pulling out the terms that drive compliance tests and cash flows. An analyst validates what comes back.
Data consolidation. Trustee reports, agent notices, and internal records pulled into a consistent foundation. Staff investigate the breaks rather than hunting for them.
Reporting. Repeatable assembly of inputs for investor and compliance reporting, with people validating the final output before it goes out.
Exception management. Discrepancies surface in a monitored queue instead of living in spreadsheet tabs and email threads.
Extraction, organization, routing, reconciliation: all preparation. Credit judgment sits with the portfolio manager, working from information that is complete and current rather than half-assembled.
Governance Comes First, Not After
In regulated fund operations, how you deploy matters as much as what you deploy.
Credit agreements, investor records, and fund data are confidential. Running that material through a shared public AI service raises immediate questions about data control, human review, and what evidence exists for compliance teams and allocators doing operational due diligence.
This is not hypothetical. The SEC’s Division of Examinations put artificial intelligence in its fiscal 2026 examination priorities, released November 17, 2025. Examiners will check whether firms have adequate policies to supervise their AI use, and whether public claims about AI capabilities match what a firm actually runs. Third-party vendor oversight appears alongside it.
Three questions follow from that: where does your AI process fund data, who reviewed the output, and what record exists. A manager without ready answers is answering an examination question badly.
Allocators expect the same straight answer. So AI in this setting needs defined boundaries, review checkpoints, escalation paths, and a record of what the automation did. An audit trail supports review. It does not move responsibility for a decision from a person to a model.
Nu-Age’s platform for the CLO and hedge fund sector runs air-gapped, privatized models rather than shared public tools. Immutable automated logging sits alongside it, producing an audit trail for SEC and NYDFS compliance.
In June 2026 the company announced expanded AI, data, and automation services for CLO managers, hedge funds, and other alternative investment firms. The scope covers workflow automation, data consolidation, analytics and reporting, and exception tracking. The privatized environment is designed to keep deal documents inside the firm’s control while logging records automated actions for later review. Those are governance controls, not regulatory approval, and they do not replace a firm’s own compliance obligations.
“The structured credit and alternative investment landscape continues to demand greater speed, precision, and operational discipline,” said Anthony Chillino, President of The Nu-Age Group. “Our expanded focus is centered on helping CLO managers and hedge fund operators improve the way information moves across the organization, reduce reliance on fragmented manual processes, and create a stronger operational foundation for decision making and growth.”
Two Published Engagements
Both results below come from The Nu-Age Group, and they come from two separate projects. The quantified figures belong to the first one only.
A hedge fund replacing its spreadsheet layer
In Nu-Age’s custom hedge fund application development case study, a fund was running core processes on manual spreadsheets and aging systems. Repetitive work ate staff time, handoffs produced errors, and reporting lagged the portfolio. Nu-Age replaced the spreadsheet layer with custom-built fund applications.
Reported results: 60% reduction in staff time on repetitive tasks, 90% faster reporting, better compliance readiness, and faster investor communication.
Worth being precise about what that is. It was a custom-application engagement, not an AI deployment, and the numbers should not be relabeled as CLO AI results.
A CLO manager rebuilding infrastructure
From Nu-Age’s managed IT practice for CLO and hedge funds: the firm restructured the infrastructure behind a multi-vehicle CLO manager’s Wall Street Office environment.
The published outcome was stronger application performance through month-end and reporting cycles, with fewer workflow disruptions during the periods when operations teams can least absorb them. That result is qualitative. No percentage, asset level, or timeline is attached to it.
Neither project was a trading algorithm. Both addressed the systems and information flows underneath fund operations, through different scopes and with separately attributed outcomes.
A Practical Sequence
Built from the capabilities Nu-Age announced in June 2026. It is a way to organize the work, not a claim that every CLO fund should follow the same path.
- Map the workflows. Document operational and reporting processes end to end, and have the people who own that work confirm the map. Automating a broken process just produces broken output faster.
- Consolidate the data. Get trustee data, agent notices, and internal records into a reconciled foundation. Fragmented source data has to be fixed before any downstream reporting can be trusted.
- Use AI for the first pass. Air-gapped models extract credit agreement terms and identify waterfall triggers. Analysts validate and approve. Documents stay inside the firm’s control.
- Automate the routing. Defined workflows replace processes stitched across files and inboxes. Final reports still require human validation.
- Track exceptions properly. Monitored queues with clear ownership instead of spreadsheet tabs and email chases. Systems flag; staff determine cause and resolution.
- Log everything. Immutable automated logging records each automated action and builds the audit trail. Compliance reviews what happened.
The goal across all six is fewer manual touchpoints, faster reporting, and the ability to add deals without adding headcount, while decisions stay with people.
Key Takeaways
- Start with operations, not trading. Document processing, consolidation, reporting, and exceptions are bounded and reviewable.
- Fitch’s “AI assisted, human decided” model separates information preparation from credit judgment.
- Governance is a prerequisite. Controlled data handling, human checkpoints, and immutable logging belong in the design, not the retrofit.
- Attribute results carefully. The quantified figures above come from a custom-application engagement; the CLO infrastructure result is qualitative.
Frequently Asked Questions
What is a collateralized loan obligation?
A securitization vehicle holding senior secured corporate loans, financed by issuing debt and equity tranches with different risk and return profiles. A manager oversees the loan portfolio, compliance tests, and payment waterfall. It differs from a collateralized debt obligation and a collateralized fund obligation, which are separate structures.
How does a CLO fund work?
It pools corporate loans and finances the portfolio by issuing tranches. Interest and principal collected from the loans get distributed according to a defined waterfall. The manager selects and monitors loans, trades within the governing documents, and tracks portfolio tests throughout the deal’s life.
How are CLO managers using AI?
Per Fitch’s May 2026 survey: mainly research support, document processing, surveillance, and workflow automation. In practice that means first-pass term extraction, data consolidation, reporting support, and exception monitoring, with people reviewing output and retaining investment judgment.
What is the “AI assisted, human decided” model?
Fitch’s phrase for an operating model where AI prepares information and flags issues while people keep authority over consequential decisions. Models extract, reconcile, organize, and monitor. Staff handle exceptions, escalation, and the credit and trading calls.
What are the risks of using public AI tools for CLO operations?
Confidentiality, control, and auditability. Sensitive credit agreements, investor records, and fund data processed outside a controlled environment create questions a firm may not be able to answer during diligence. Firms need data-handling boundaries, human review, documented controls, and audit evidence.
The Nu-Age Group has built secure infrastructure for regulated firms for nearly three decades, with AI, data, and automation services aimed at CLO managers, hedge funds, and alternative investment firms. To discuss a workflow assessment, visit financial services technology and security, call (866) 640-3999, or email sales@thenuagegroup.us.
