A credit score answers one question well and every other question badly. It tells you the probability of failure across a population of similar entities. It does not tell you which counterparty is about to stop paying, which concentration will hurt first, or what the terms ought to be.
That is not a criticism of scores. It is a statement about how much a single number can carry. The score is the entry point to the decision. The decision itself needs a profile built around it.
Why one number holds one answer
A score compresses a long list of behaviors into a rank. That compression is what makes it useful at portfolio scale and what makes it useless at the level of a single decision. Two counterparties can share a score and share almost nothing else, because the score discards the reasons behind the number.
The reasons are where the decision lives. A strong score sitting on concentrated revenue from one customer, thin liquidity and a lender that has just taken a secured position is not the same risk as a strong score sitting on diversified revenue and improving margin. The score cannot see the difference. The committee can, if it is given the profile.
What a profile contains that a score cannot
Identity. Whether the entity is who it claims to be, who actually stands behind it, and whether the person signing has the authority to sign. A false identity invalidates every signal downstream of it.
Financial position and trajectory. Not just the ratios, but their direction across four quarters, and whether the movement is consistent with the industry the counterparty operates in.
Payment behavior. Days beyond terms, the consistency of the pattern, and whether the delay is widening across invoices. This is the least flattering and most predictive evidence available.
Counterparty and procurement concentration. How much of your exposure sits with this entity, how much of their revenue depends on you, and where their own supplier relationships create single points of failure that become your problem.
Ownership and legal events. Filing changes, secured positions, officer departures and ownership transfers. These are the events that move risk before the financial statements do.
From descriptive to prescriptive
Descriptive analytics tell you what happened. Diagnostic analytics explain why it happened. Predictive analytics estimate what comes next. Each one is a layer, and each one is incomplete on its own.
The layer that changes a decision is prescriptive. It answers a question the other three cannot: given everything we know, what should we do about this counterparty, and what would have to become true for that answer to change?
A profile built this way produces four outputs that a committee can act on: a limit, a term, a security position and a review cadence. Each one traces back to evidence, which is what makes it defensible when the decision is challenged six months later.
The procurement dimension
Counterparty risk does not stop at the receivable. The same entity appearing in your supplier base creates financial exposure in a different direction, because a supplier that fails disrupts revenue as surely as a customer that defaults.
This is where the network view earns its place. Mapping the same entity across customers, prospects, suppliers and counterparties turns four separate reviews into one risk position, and it surfaces the overlap that no single system reports.
What changes in the decision
The score still does its job. It sorts the portfolio, it sets monitoring frequency, and it flags where the profile work should concentrate first.
What changes is the second half of the decision. Instead of approving or declining on a threshold, the committee sets conditions: this limit at this term, with this security, reviewed on this cadence, and with this trigger for re opening the file. That is a risk position rather than a verdict, and it is the difference between managing exposure and reacting to it.
Internal Knowledge + External Insights = Competitive Intelligence. That is your Decision Intelligence Framework.

