AI in lending: how artificial intelligence opens the door for SMBs

Lending services are a critical and often underappreciated part of the business landscape. Especially for small businesses, working capital can be the difference between fortune and failure. So lenders really can be the unsung heroes of our modern economy. 

But traditionally, access to financing has been limited, particularly for those SMBs. The biggest issues are:

• Time, with loans and factoring agreements regularly taking more months to secure
• Administration, with the documents and due diligence required simply overwhelming some companies
• Cost, as lenders try to maintain profit margins in the face of the previous two factors 

This has been the status quo for some time. But artificial intelligence is here to change this dynamic. Suddenly, lending can be faster, simpler, and cheaper than ever before. It can also be safer and more flexible. 

While that might sound like a fever dream, it’s very much the reality. In this article, we’ll share exactly how AI enhances and updates lending services, and what this means both for lenders and borrowers. 

How artificial intelligence enhances lending services

Lending is a complex financial service. Vetting and approving new clients requires lots of due diligence and document verification, and there’s also the administrative cost of creating accounts and managing a growing client portfolio. 

But thanks to AI and machine learning, the labour required is quickly diminishing. 

We’ll look at specific use cases in a moment. But because AI can digest and analyse huge amounts of data and piles of documents in seconds, that administrative burden comes way down. 

First, this has a tangible effect on lenders' bottom lines. You need fewer staff and less time to review documents and complete KYCs, and operational costs plummet. This includes non-standardised documents which were previously hard to automate: fiscal statements, proofs of payment, and unorthodox customer contracts. 

It previously took a human agent up to an hour to review a new client file — really just to get the process started. And the more unusual or atypical these documents were, the longer the processing time. 

At Defacto, we can do this in just 27 seconds per client. And we can process endless new clients simultaneously. 

So lenders’ profit margins go up. And they can also take on more clients, from a more diverse range of profiles. Which is great news for lenders — it’s now easier to do more business. 

But it’s also great for previously overlooked industries, and for SMBs in particular. No file (or customer) is too small or simple. Every application can be considered on its own merits, and a line of credit can be extended almost instantly. 

Finally, the quality of lending services can improve. Compliance checks are not only faster, but more complete. In the past, you might have had time to review 10-20% of documents in detail. Today, AI tools can review 100%. Instances of fraud or misconduct are far less likely to slip through. 

And you provide a faster, easier service for new clients. No business owner wants to wait days or weeks to access funds. Now, they don’t have to. 

How AI optimises lending from end to end

Let’s look at a few of the specific ways that AI improves the classic lending workflow. We can broadly bucket these into upgrades to the customer-facing user experience, and back-end advantages for providers.
‍

User experience improvements‍

• Document verification and faster KYBs. AI tools can classify documents and instantly determine whether it’s real and legitimate. They can also check your records for duplicates and flag attempts to open multiple accounts.

• Customer onboarding. Actually creating the customer account and getting them up to speed with your service can also be largely or entirely automated. And it can still be a delightful customer experience, made even better because of how quickly and effortlessly it’s completed.
  

• Customer service and support. Where customers want a quick, direct answer to a particular question, AI tools are ideal. They instantly review a file and find the best response in real time. This takes much of the burden off supporting smaller clients in your portfolio.
  
  And for more complex queries, an AI copilot can deliver all the context and background information a human agent needs quickly. They have examples of similar cases and sample responses at their fingertips, without having to search.
  
  
• Personalised financing recommendations. AI can help build tailored financing services based on the industry, business model, and behaviour. They can also spot lookalike customers and suggest options that have worked well in other cases. This takes you beyond standardised packages into bespoke, personalised funding at scale.  

Back-end operations upgrades

• Risk analysis and fraud detection. Lending requires companies to verify their identity, income, registration, and much more. The faster you can verify these documents, the more applications you can process. AI can also notify you if customer documents are approaching an expiry date, or if regulations require additional information.
  
  
• Underwriting and loan origination. On top of verifying that documents are real, AI can extract the data from income and bank statements to ensure that potential clients have assets and a solid credit history. It can also search public records and registers to ensure prospects haven’t been flagged for misconduct in the past.
  
  And then creating the loan and transferring funds can be almost instant. 
  ‍

• Servicing and credit management. Think of AI tools as a copilot to help you manage a portfolio of clients, each with their own repayment schedules and possible compliance issues. The tools can flag problems, highlight your most and least profitable customers, and send notifications to clients that payments are due (or overdue).
  
  
• Refinancing. Good AI software can even spot spend trends among your customers, and identify opportunities for them to refinance. This could include seasonal financing during peak times, or where they’ve used financing to great effect and another round could help them grow more quickly. 

‍

And these aren’t just wishful thinking or hypotheticals. At Defacto, we use artificial intelligence for every task on that list — and plenty more. 

The positive effects of AI automation in lending

Adding AI to lending processes isn’t just a flashy attempt to modernise or keep up with the times. There are tangible, real-world benefits for all parties.

Businesses can better optimise their cash flows

Businesses need convenience and speed, which isn’t always guaranteed with traditional SMB lending. Companies with urgent cash needs can quickly onboard and access financing before it’s too late.

On the positive side, businesses can take advantage of hot opportunities. They can quickly finance inventory or hire new staff to meet increased demand.

This is great news for the wider economy. We don’t want lenders to operate as the ambulance at the bottom of the cliff, only stepping in when companies are on the brink of disaster. Financing should be a growth lever, not a parachute. 

SMBs (in particular) get new access to financing

Because financing processes can be time consuming and administratively intense, too many small businesses are left out. It’s either not profitable for lenders to take on these clients, or the time frame is too long and the SMB can’t afford to wait. 

More automation and AI in lending reduces this administrative burden, so SMBs can get the same opportunities as everyone else. 

Lenders (including banks) serve more customers, better

On the flipside of the equation, lenders have more incentive to serve a wider range of customers, including those same SMBs. Embedded credit services cause administrative costs to plummet. So even clients with a narrower potential for profit can be profitable nonetheless. 

There’s also the possibility to explore more compliance-intensive business models and industries. If the KYC process was previously too long and arduous to make it worthwhile, that’s no longer the case. 

Fintechs can expand with low overheads

Finally, AI lending platforms themselves benefit from being part of the wider financial ecosystem. Startups like Defacto aren’t built to replace banks or lending institutions. We want to integrate and provide the tools these establishments need to play their vital role in the economy. 

Which of course has big benefits for fintechs. We can grow through these integrations without over-investing in go-to-market strategy. By keeping sales, marketing, and operating costs as low as possible, we can stay focused on what we do best: building the new wave of lending technology.

More platforms can incorporate lending

We’re huge proponents of lending-as-a-service (Laas) platforms. These let virtually any platform build lending into their own offerings, without building an entire lending product of their own. 

While LaaS integrations existed before the recent AI boom, technological progress makes it easier and lower cost to add them to an existing product. As lending becomes cheaper and the administrative burden decreases, we’ll see even more use cases for lending across the tech sector. 

What are the limits for AI in lending?

We firmly believe that AI is already changing the nature of lending in good and exciting ways. And there are bound to be countless other use cases and advantages nobody has thought of yet. 

But despite being bullish, we must have guardrails. Compliance and data sensitivity are crucial in financial services, and we can’t afford to jeopardise them. 

A few things lenders shouldn’t do with AI:

‍

‍

• Send confidential data to public APIs. Customer data must never be sent to public APIs, including AI GPTs. You should want to move fast, but safety first in all things finance. 
• Put out-of-the-box models into production. The beauty of large learning models is their capacity to evolve with training. You can’t just plug a model into your production environment with no testing or tweaks. It’s not safe, and bound to produce a bad user experience. 
• Let AI fly solo. Even if AI can do a lot of heavy lifting, you still need human compliance and legal experts to ensure that the tools follow the rules. Your copilot needs a wing(wo)man in place. 

These tools are workhorses, there to follow instructions and lend a hand. But they’re not the final decision makers, and can’t replace industry expertise and legal knowledge completely. 

As the saying goes: “garbage in, garbage out.” Without well-structured data and plenty of fine tuning along the way, you’ll never see the service enhancements we detailed above. 

AI brings a much-needed upgrade to lending services

Access to funding is vital for businesses of all sizes, but particularly small businesses. And while business loans and lending have been relied on forever, we’re past due for an update. 

This is exactly what artificial intelligence provides. We can now give small businesses — and all businesses — the financing they need, when they need it most. 

AI makes lending:

• Faster for both the financial institution and the customer. 
• Cheaper by reducing the administrative burden all around
• Fairer and more inclusive, by reducing barriers and letting institutions serve a broader range of businesses
• More compliant because we can do due diligence at scale

This new technology is a huge win/win for lenders and their customers. Talk to us to learn how we use AI to deliver fast, fair, flexible financing at Defacto. And how you can put the same tools to use in your own services. 

Interested in digging into more of our AI use cases? Watch the recording of our panel at Raise Summit where Marco, our co-founder and CTO, discusses foundational models alongside leaders from Deepset, Jina AI, Raive, and Sifted. 
‍