The Book You Can’t See: Portfolio Monitoring Is the Next Underwriting Frontier

By Steve Iskander, Founder and CEO, Intrepid

In this article

• The asymmetry between origination and monitoring
• Signals that arrive before the breach
• Why the regulatory backdrop matters now
• Building monitoring that does not depend on memory

We spend enormous effort deciding whether to make a loan and comparatively little on loan portfolio monitoring after close.

Think about the resourcing. Underwriting gets a team, a committee, a memo, a policy document, and increasingly a model. Monitoring often gets a quarterly reporting package, a spreadsheet of covenant tests, and whoever has time.

Then we are surprised when a problem shows up late.

This is on my mind because the environment has changed underneath the practice. The Financial Stability Board raised valuation opacity and monitoring gaps in private credit this year, and US regulators have sharpened their attention on how marks and ratings get produced. Independent monitoring specialists have been making a quieter point for a while: the data anomalies that precede a covenant breach are often visible well before the breach is reported.

That is the part worth sitting with. The signal usually exists. The surveillance cadence is not fast enough to catch it.

A borrower’s receivables slow in month one. Payroll timing shifts in month two. A concentration builds through month three. On a quarterly package all three arrive at once, after the quarter closed, described in a narrative the borrower wrote. On live data each one is a flag in the week it happens, while the conversation is still a conversation and not a workout.

None of that removes judgment. A slow receivable can mean a new enterprise customer with ninety-day terms, which is good news. Anomalies are not verdicts. The point is that you should see them while you can still ask.

The infrastructure argument here is the one I have been making all month, applied to the second half of the credit lifecycle. If your data is structured and live for underwriting, monitoring is nearly free. If it is not, monitoring is a manual process competing with new deals for the same people, and new deals win every time.

That is why we built covenant testing into the platform rather than beside it. Same pipeline, same live data, applied continuously after close. Outcomes vary by portfolio and covenant structure.

Underwriting is a decision. Credit is a relationship you maintain.

For the portfolio teams here: how long between a borrower’s numbers turning and your team knowing about it?

The asymmetry between origination and loan portfolio monitoring

Lenders pour resources into the decision to make a loan and comparatively little into watching what happens after. Origination gets a team, a committee, a policy, and increasingly a model. Monitoring often gets a quarterly reporting package and whoever has time. The result is predictable: problems surface late, not because anyone was careless, but because the surveillance cadence was built for periodic reporting while businesses deteriorate continuously.

Signals that arrive before the breach

The useful insight from monitoring specialists is that the data anomalies preceding a covenant breach are often visible well before the breach is reported. Receivables slow. Payroll timing shifts. A single customer becomes a larger share of revenue. Reporting drifts later and thinner. On a quarterly package these arrive together, after the fact, wrapped in a narrative the borrower wrote. On live data, each one is a flag in the week it appears, while the situation is still a conversation rather than a workout. None of these signals is a verdict; each is a reason to ask a question early.

Why the regulatory backdrop matters now

This is not only an operational question. In 2026 the Financial Stability Board flagged valuation opacity and monitoring gaps in private credit, and United States regulators sharpened their focus on how marks and ratings are produced. The direction of that scrutiny rewards lenders who can demonstrate timely, data-driven visibility into their books. Continuous monitoring is becoming both a risk-management advantage and a way to meet a rising bar for transparency.

Building loan portfolio monitoring that does not depend on memory

The goal is surveillance that does not rely on someone remembering to look. If the underlying data is live and structured, the system can watch continuously and escalate to a human only when something moves, a different economic model than assigning people to read quarterly packages. Covenant tests run automatically. Anomalies surface as they appear. A performing loan that is quietly drifting gets attention before the drift becomes a problem. This matters most for the credits that look healthy, because attention naturally flows to new deals and to names already in trouble, leaving the quiet, performing borrower unwatched. That borrower is precisely where a surprise tends to originate. Continuous monitoring closes that blind spot without burying the portfolio team in more reporting. Underwriting is the decision to lend. Monitoring is how you find out whether you were right, and catching a problem early, while it is still a conversation, is almost always cheaper than discovering it in a workout months later. The lenders who invest here are not adding paranoia. They are moving the moment of discovery earlier, which is where nearly all of the value in monitoring lives.

Related reading

→ Why underwriting speed is a data problem

→ Why lending data infrastructure is the moat

Frequently asked questions

What is continuous covenant monitoring?

Continuous covenant monitoring is testing a borrower’s compliance with loan covenants on an ongoing basis using live data, rather than checking only at quarterly reporting intervals. It lets a lender see a potential breach in the week it develops instead of the quarter it is reported, while the situation is still a conversation rather than a workout.

Why are covenant breaches detected late?

Breaches are often detected late because monitoring is built around periodic reporting packages and manual, spreadsheet-based tracking. Businesses do not deteriorate on a quarterly schedule, so signals that emerge mid-quarter are not seen until the package arrives, sometimes months later.

What early-warning signals appear in borrower data?

Common early signals include slowing receivables, shifts in payroll timing, a build in customer concentration, and drift in the timeliness or completeness of reporting. Independent monitoring specialists note that these subtle anomalies often appear before a covenant breach is formally reported. They are prompts for inquiry, not conclusions.

What did the FSB say about private credit transparency?

In 2026 the Financial Stability Board flagged concerns about valuation opacity and monitoring gaps in private credit, and US regulators sharpened their attention on how marks and ratings are produced. The direction of that scrutiny favors lenders who can demonstrate timely, data-driven visibility into their books.

How is continuous monitoring different from quarterly reporting?

Quarterly reporting looks backward at set intervals, so mid-quarter signals arrive late and bundled together. Continuous monitoring watches live data as it changes, surfacing anomalies in the week they appear. The difference is often the difference between an early conversation and a late markdown.

Sources: Financial Stability Board

Intrepid turns live financial data into loan portfolio monitoring that surfaces problems as conversations, not workouts. See how at intrepidfinance.io.

Published by Intrepid. Democratizing Access to Capital. intrepidfinance.io