By Richard Robinson, CEO and founder of Robin AI
When it comes to credit lending, the early bird catches the worm. The world of private debt investment is fraught and highly competitive, where those with the fastest fingers are best placed to capitalise on the most promising and lucrative new opportunities.
Whether they’re raising a fund or looking to deploy the capital within it, speed is everything to global asset managers. They know that if they blink (let alone sleep), they risk forgoing the opportunity to a competitor.
However, for the professionals working on the legal side of deal-making – not exactly shirkers when it comes to hard work and long hours – the priority is accuracy, not speed. From completing vast due diligence questionnaires (DDQs) to creating, reviewing and amending complex NDAs, they will take as long as it takes to get the job done properly. Indeed, given the vast sums being invested, they can ill afford to make mistakes, even if this level of rigour and diligence risks causing delays or even missed opportunities.
Consequently, the legal side of credit lending—in which human domain expertise has always been sacrosanct—has gone untouched by the mass digitalisation we have witnessed over the past 15 years. While software, analytics and algorithms have transformed essential business processes in other departments, legal professionals still rely heavily on word processing and spreadsheets to carry out their duties.
The rise of the large language models changes everything
Since late 2022, the advent of generative artificial intelligence (AI), specifically, the creation of large language models capable of seemingly human-like cognition in response to basic text instructions, has fundamentally changed the technology landscape.
AI promises to transform knowledge work like never before and undoubtedly holds great promise for the entire credit lending industry. One notable example is the potential for radically improved risk modelling, enabling asset managers to choose their investments more wisely and uncover previously overlooked opportunities.
However, the greatest near-term advantage lies in transforming the efficiency of credit lenders’ everyday legal processes. By leveraging large language models and machine learning, AI can accelerate tasks such as turning around due diligence questionnaires (DDQs) and NDAs from days or weeks to a matter of hours. Implemented correctly, it offers a step-change in the speed at which legal professionals can operate without any compromise to the accuracy of their work.
AI will help close funds faster
Private debt fund formation involves a massive amount of due diligence. Global asset management companies will likely need to complete DDQs from every prospective investor. The DDQ covers everything from previous fund performance to the profiles of the fund managers involved and, of course, the fees involved in the partnership.
These are long, complex documents, requiring input from across the firm. Completing them takes up a huge amount of time for investor relations teams, finance teams, legal & compliance teams and beyond; it’s not uncommon for turnaround times of 10 days for more complex DDQs. What’s more, these documents are not just needed at the outset of a new relationship, they’re an ongoing investor requirement, often delivered to an agreed cadence such as every six months or every quarter.
Speed is of the essence; any delays risk investors opting to put their capital elsewhere. Yet this laborious, repetitive work takes investor relations teams and their legal advisors weeks to complete.
By contrast, an LLM trained to understand the language of private debt assets and credit lending can significantly accelerate DDQ completion by looking at historic questions and retrieving relevant information from across a whole suite of documentation within seconds. All the human knowledge workers must do is verify the information retrieved by the AI, and the process is completed in a fraction of the time normally taken.
Once the DDQs are out of the way, AI can play a vital role in helping close the fund by instantly parsing through the 100s of Limited Partner Agreements (LPAs), side letters and complex obligations demanded by investors to help credit lenders’ legal teams ensure consistency and conformity across the board.
This is hugely significant work. Each LP will have non-negotiable requirements about how and where their capital can be invested, and the terms involved. Legal professionals currently rely on giant spreadsheets to house these myriad clauses, which not only have to be squared at the fund formation stage but also carefully referenced prior to every subsequent capital deployment. By using AI to do the cross-referencing for them, they can deliver the appropriate deal terms faster than ever previously thought possible.
AI enables a crucial first-mover advantage
While there is no shortage of companies seeking credit lending, the best private debt deals are highly sought after, making first-mover advantage a necessity for asset managers looking to maximise fund returns.
Competition in this market landscape is intense. Contracts are a necessary but time-consuming step in the process; any delays could put fast-moving deals at risk. When minutes and hours make all the difference, the teams that can get the routine contracts out the way the most quickly have a significant advantage.
Again, AI holds transformative potential for legal teams as they look to accelerate the NDA process. They can train LLMs on their corporate playbook—i.e., the standard NDA terms used for certain deal types—and use AI to instantly highlight where each company’s proposed NDA differs from their approach, allowing them to make edits and reach a mutual agreement faster.
By streamlining routine work like NDA reviews, AI can give asset managers a first-mover advantage and free up their busy legal teams to focus on progressing the details of the deal.
There are limitations to what AI can do – and lawyers must remain in the loop
AI is constantly evolving, and barely a month goes by without one of the big LLM providers unveiling a new model professing greater cognitive capabilities.
Yet it would be a mistake for credit lenders to over-rely on these technologies, or worse, to assume that AI can replace human domain expertise within their businesses.
Asset managers shouldn’t even consider using generic LLMs like ChatGPT to parse their investor data or make recommendations on the specific legalese of their LPAs or corporate NDAs. These AIs haven’t been trained on this type of data, and attempting to feed massive amounts of sensitive information into them is both a security risk and a massive privacy concern.
Instead, they must look to purpose-built legal AI tools that can harness the power of LLMs while guaranteeing the security and confidentiality of their data – i.e. that nothing will be relayed back to the big AI providers.
Likewise, all LLMs are capable of hallucination – a problem that is already well-documented in the legal sector. Credit lenders need to focus on applying AI where it can be a legitimate aid to lawyers-in-the-loop, namely, automatically surfacing information for professionals to verify, rather than attempting to automate entire processes such as DDQs, LPAs and NDAs.
The future of credit lending will be AI-driven
Most global asset managers rely on lean in-house legal teams – flanked by massively expensive external support – to create new funds and deploy fund capital effectively.
To date, many of these credit lenders have been willing to accept increased legal costs as a necessary consequence of their desire to move faster and secure the best investors and the best investment opportunities.
By synthesising large amounts of documentation, AI will quickly have a similar impact on the legal profession as Excel did on accounting. With the right legal AI tools, asset managers can eliminate the speed/cost trade-off and avoid the need for unnecessary third-party expenses, enabling their in-house teams to focus resources on the work that will directly impact the bottom line. We’re on the brink of an AI revolution in credit lending, and it all starts in legal.
