For lenders, accurately assessing borrower risk is one of the most critical and costly parts of the job. Traditional credit scoring models often fall short, especially for small and medium-sized businesses (SMBs). These often have limited credit histories or complex financial profiles. And you need assessments to be quick and concise, as SMBs are usually deemed low-profit borrowers.

As a result, you’re forced to rely on incomplete data, outdated methods, or time-consuming manual reviews. And more often than not, SMBs are left without the vital financing they need. 

The two problems here are the kinds of data most lenders rely on, and the time and effort required to go further. And we’re proud to say that we’ve found a better—and faster—way of assessing prospective borrowers. 

We call it “node-based scoring,” and we’re excited to show you how it works. 

The trouble with the status quo

Credit assessments have already come a long way in recent years. Automated underwriting lets lenders evaluate and onboard new borrowers faster than ever, with more security checks along the way. 

This is great. But it’s still built on incomplete information. 

Businesses don’t operate in isolation. They’re built on networks of clients, suppliers, and partners, forming a web of relationships. But  the typical credit scoring system doesn’t reflect this. 

Credit underwriting today

In typical lending, you assess the creditworthiness of prospective borrowers using publicly available information and data, including:

• Publicly available information. The company’s address, age, and listed directors. Plus the equity invested or capital raised.
• Financial data. Balance sheets and other core financial statements.
• Transaction data. Invoices, receivables, and bank transfers via open banking APIs and cloud accounting platforms. 

In certain countries, lenders also rely on credit scores from bureaus like Equifax or Creditsafe. These mainly use the same publicly available information to determine a score for each business.

And that’s usually where it ends. The lender has a picture of a potential borrower at a certain moment in time, based on relatively basic information.

But crucially, while it looks closely at the risks posed by a company’s own directors and operations, it ignores the partners, suppliers, and customers they work with.

“Node-based” scoring

Each company has its own ecosystem, with links to suppliers and customers reaching out like a spider’s web, a neural network, or whichever metaphor you prefer. And any one of these links—or “nodes”—is a possible risk or asset to that company.

(We call these supplier or customer nodes “counterparties”.)

A potential borrower may look good on paper—strong financials, plenty of equity, and no obvious red flags in its directors or address. But suppose 75% of their stock comes from a partner that has defaulted on loans in the past, or has recently declared insolvency. 

Suddenly, that attractive borrower looks risky—if they can’t find a new supplier quickly, they likely go out of business. 

With enough due diligence and examination of past transactions, you might have caught this. But relying only on traditional underwriting data, you wouldn’t have. 

And you certainly wouldn’t be able to analyze each of the dozens of key suppliers in their supply chain. 

Mapping the business ecosystem

Node-based scoring is essentially a map of every borrower’s network. By pulling together vast amounts of public and private data—everything from open banking transactions to ERP (enterprise resource planning) records—we can build a detailed graph of who’s working with whom. 

Node-based scoring uses a range of data points, including:

• Publicly available information and finance statements. We look at company addresses and directors’ histories. Are either linked to other companies which proved risky in the past? And we also assess key financial statements. This is essentially “traditional” credit scoring.
  
  
• Transactions: Invoices, open banking, and accounting APIs bank transactions and invoices via API and open banking. Most underwriting automation does this already, too. But ours goes further.
  
  LLM models and our own proprietary algorithms can extract the counterparties they’re dealing with. These are often not given directly by open banking partners, but we can identify counterparties through transaction information. We can then score these counterparties the same way we would do for a new Defacto client.
   
• Defacto history. We’re lenders too, and in many cases we already have history with SMB suppliers. What do we know about these counterparties, and does a borrower working with them raise any concerns?
  

With all these data points, we can:

1. Assign a risk score to each individual counterparty; and
2. Construct an overall financial health score for each borrower, based on all the counterparties they rely on. 

It’s similar to the domain rankings used in search engine optimization: Google knows how authoritative one website is based on the quality of websites linking back to it (its “backlink profile.”)

By mapping out and evaluating all counterparties, you get a far more holistic and complete assessment of the level of risk any new borrower carries. 

Why this is exciting for lenders

Node-based scoring brings very real benefits for lenders. 

1. Easier fraud detection

Fraud remains one of the biggest risks in lending, particularly when borrowers intentionally create fake or circular transactions with related businesses to inflate their credibility. Traditional models that only look at individual company metrics often miss these fraudulent networks, leaving lenders exposed to bad debt and reputational damage.

The most obvious benefit of node-based scoring is to reduce the risk of default or fraud in the lending process. A flagging mechanism highlights potential node issues with prospective borrowers. 

If they’re linked to insolvent suppliers or have risky directors, we’ll flag this immediately.

2. Fewer default risks

Lenders also get a broader algorithmic picture (or scorecard) of the financial health of any business. If their node ecosystem is lower risk overall, this suggests the business itself is lower risk. 

The beauty is that this evaluation is totally independent of whether or not a business itself is behaving in risky ways. By identifying both strong and weak nodes in the network, lenders can make better decisions, reduce fraud, and extend more loans safely.

3. Better ongoing monitoring

Node-based scoring doesn’t end at the loan initiation phase. Lenders can monitor financial health across their entire portfolio on an ongoing basis, particularly where they have open credit lines and recurring customers. 

For example, if a company declares insolvency, you can see all of your borrowers in the portfolio who are linked in some way to that business. Borrowers who were previously low risk may have a new risk profile if a key supplier’s status changes. 

4. New growth opportunities

On the positive side, you can spot opportunities to strengthen relationships with reliable borrowers. A borrower who works consistently with trustworthy partners, or whose customers continue to pay on time without fail, is the kind of client you want in your portfolio. 

Lenders can more confidently offer higher limits, extended credit lines, or additional financial products to their most trustworthy clients.

5. The ability to offer new services

You can also use node-based scoring to deliver new and better services as lenders. If a supplier becomes risky, you can alert customers to the possible issues they face in future.

You can also alert borrowers instantly if a supplier has declared insolvency. This gives them more time to reach out to the counterparty, or to immediately find a new supplier. 

When a customer submits an invoice for financing, you can quickly scan that customer or supplier and provide advice or warnings if things look suspicious. You can also choose not to lend money against that risky invoice, of course.

The upshot is happier, healthier customers, and the potential for new paid lending services. 

6. A much richer data pool

Many SMBs don’t have long credit histories, audited financials, or formal collateral, making it difficult for lenders to assess their risk. Even when data is available, it’s often fragmented across banking records, tax documents, and accounting systems. Without a unified view, lenders struggle to build a clear, accurate picture of the borrower’s financial health.

Even in cases where companies have common names, Defacto’s large language model (LLM) can correctly identify the right entity by factoring in industry context and location. With the coming wave of e-invoicing, the system will gain access to even richer, certified data, making the map even more precise.

7. Automation at scale

To fill the gaps in traditional credit data, lenders often resort to manual processes: reviewing bank statements, analyzing invoices, or even calling suppliers for references. This detective work takes time, adds operational costs, and slows down the decision-making process—especially when dealing with a large number of applications.

And while credit scoring is now largely automated, most lenders stop at easy-to-access public data and basic financial records. Node-based scoring creates a much more accurate, complete borrower profile with no additional work. Lenders no longer need to act like detectives, piecing together clues from disparate sources. You get an instant, high-resolution picture of a borrower’s credibility.

Faster, fairer underwriting for more SMBs

Traditional scoring often overlooks promising small businesses simply because they lack a long credit history or collateral. But node-based scoring opens lending opportunities for much smaller, less traditionally “profitable” borrowers. 

Even a tiny company can shine in the right ecosystem. And because the scoring system is automated and instant, you can fairly evaluate every potential borrower.

The primary goal is to make lending safer and more efficient for lenders. But the secondary effect is that more SMBs will have access to the funds they so desperately need. Cash flow worries and access to capital are among the most fundamental reasons small businesses go under. 

Today, they’re too often locked out of financing because of the costs (and limited rewards) associated with serving them. Many SMBs with excellent clients or strong payment records get overlooked simply because they are young or operate in industries without well-established benchmarks.

Node-based scoring is a critical tool in changing this dynamic. 

See the model in action

Defacto’s node-based scoring model is more than just a new tool—it’s a new way of thinking about risk, trust, and opportunity in the world of SMB lending. Leveraging the power of business networks, lenders can make faster, more confident decisions, uncover hidden fraud risks, and unlock growth opportunities with their best borrowers.

Whether you’re a bank, a credit union, or a SaaS platform looking to offer embedded finance, this technology can transform how you assess and serve your customers.

Interested in seeing the model in action? Contact us to learn how this breakthrough solution can help your lending business reduce risk, increase efficiency, and build stronger, more profitable relationships.