On this page, learn about investment banking automation topics including:
Investment banking companies deal with all sorts of documents on a daily basis, from customer onboarding forms to ISDA Master Agreements, background for underwriting services, SEC 10-Ks, complex merger & acquisition agreements, and more. Automating the processing of these documents – literally millions of them – is crucial to increasing efficiency, productivity and profitability.
Banks have tried to address the problem with automation tools that use keywords, rules-based methods and templates. But they’ve had limited success for a simple reason: such tools can’t handle unstructured content and data. Leading-edge investment banks are finding the solution is intelligent document processing systems that use artificial intelligence technology. Such tools can comprehend unstructured documents and process them just as an experienced employee would – making the tools an automation game-changer that bring significant, immediate value.
More revenue with the same headcount
Streamline and codify processes
Automation systems that rely on robotic process automation, keywords, rules or templates are no match for unstructured content because the documents involved are too varied in nature. Such approaches succeed only when the same data is in the same place in each document. Only then can the template identify which data to extract.
Using such an approach with unstructured content would require hundreds or thousands of templates to account for every possible document permutation – a nearly impossible and prohibitively expensive proposition. Consider mergers and acquisitions. M&As require reams of documents from each company, from balance sheets and tax returns to letters of intent, purchase & sale agreements, escrow agreements and lots more. It takes enormous amounts of time to pore over these documents looking for the small tidbits that may represent red flags – or highly positive signs. You would have to spend millions paying consultants to come up with templates or write rules to find these tidbits in every possible document type you may encounter. Even if you did, as soon as a new document type came along, or an existing document format changed, the automation would break down. Money wasted.
Intelligent document processing technology requires no such rules or templates. Because it uses AI technology, IDP solutions can “read” and comprehend the context behind documents, images and videos just as your employees do. Beginning from a large base of labeled data points to provide context, IDP platforms enable customers to build custom models that address their specific requirements.
The Indico Unstructured Data Platform, for example, is based on more than 500 million labeled data points. The AI concept of transfer learning enables users to draw on that database to create models that automate processes involving most any sort of unstructured content, including documents, images and video. It takes only a few hours and around 200 documents to train a model to an accuracy rating of around 95%.
Even better, no data scientists are required. Whether the process involves M&A activity, underwriting, equity research, or trading activity, it’s the business professionals who best understand the processes who use Indico Data tools to build models – so-called citizen data scientists.
A primary function of an investment bank is underwriting, which involves assessing the risk of an appropriate price for securities, often related to an initial public offering (IPO) or the sale of a private company to investors. It’s a painstaking process that involves assessing numerous financial documents related to the performance of not only the company seeking an IPO or sale, but those of its competitors, in order to appropriately assess market risks. All these documents, many of them unstructured, make underwriting an investment bank process that’s ripe for automation.
Now that the LIBOR interest rate benchmark has been phased out, investment banks need to adjust the terms underpinning any loans that reference LIBOR. That means they must first find all those loans, then review them to find terms that need updating. For humans, that’s a laborious process to say the least. On the other hand, it’s a straightforward matter to train an intelligent document processing model to find loans that reference LIBOR, then examine each to identify relevant terms and data. In short order, an investment bank could identify not only all the loans that need updating, but the exact terms in each that require attention. (To learn more, reach our blog post, “Don’t Labor over LIBOR’: Meet the Looming Deadline with Intelligent Automation.”
Another primary investment bank function is providing equities research, to help clients make sound investment decisions. Each securities analyst examines SEC reports and other documents in an effort to understand the financial health of the companies for which they’re responsible. That involves reading quarterly 10-Q and annual 10-K forms looking for actionable data, then pulling it from the reports and entering it into spreadsheets. An effective intelligent document processing tool could take on the job for them, freeing up the analysts’ time to actually analyze the results. (See the blog post: “Bringing Intelligent Process Automation to Financial Document Analysis.”)
Processing over-the-counter derivatives transactions requires examining the ISDA Master Agreement that defines the terms between the two parties involved in the trade. It’s a herculean task, given the ISDA document weighs in at 28 pages and is really just a template; different variables will apply to each transaction. It can easily take a human 2 hours to process a single one and large banks may process hundreds of thousands per year – making the processing of ISDA agreements a prime candidate for intelligent automation in investment banking. (For more, see the post: “Process Automation Comes to ISDA Master Agreements.”)
Investment banks in the U.S. must comply with various regulations intended to detect money laundering. In practice, that means collecting numerous documents from clients intended to prove they’re legitimate and ongoing monitoring for any negative news about clients. Traditionally these were manual processes, but today tools like Indico Data’s IDP platform can automate large portions of anti-money laundering programs. (Read more about how automation can help in anti-money laundering efforts in this blog post.)
Chances are any investment bank that’s interested in process automation already has experience with robotic process automation, and perhaps significant investments in the technology. While it’s true that RPA cannot, by itself, process unstructured documents and data, it can complement an intelligent document processing tool to get the job done.
RPA is effective at automating processes involving highly structured content and data, and at automating tasks that involve repeating the same steps or keystrokes over and over. In such instances, RPA will not only relieve employees of mundane, labor-intensive tasks, but will likely perform them more accurately, because computers don’t get tired.
Intelligent document processing, with its ability to automate processes that include unstructured data, is a good complement to RPA, helping investment banks automate processes end-to-end.
Consider the equities research example. An intelligent document processing solution can “read” the unstructured data contained in 10-Q and 10-K forms, pull out pertinent data, and convert it to a structured format. The IDP platform could then hand off the now-structured data to an RPA tool, which would automate the process of inputting it into a spreadsheet or other downstream processing system.
It’s a misconception, however, to think process automation will completely eliminate the need for humans in many jobs. From our experience with investment banking clients, the more common scenario is companies automate a large percentage of a given process – typically the most repetitive and mundane portions, which employees are happy to give up. Employees are then free to focus on the exceptions – the instances that require their expertise and intervention. (With an IDP solution such as Indico Data’s, such interventions can often be used to help train the model to be increasingly more effective.)
In fact, only about 5% of occupations can be fully automated with current technologies, according to a recent report by the McKinsey Global Institute. It supports the idea that portions of jobs will be automated – about 30% of the activities in 60% of all occupations, McKinsey estimates.
In the investment banking world, that means you’ll be freeing up your employees for more valuable, strategic and profitable work.
In this era of digital transformation and FinTech, investment banks need to take definitive steps to drive efficiency. The Indico Unstructured Data Platform is one such transformative solution that delivers a rapid return on investment and quantifiable benefits, including:
Before you automate a process, you must agree on how it’s performed. The codification exercise means you’ll capture knowledge on process procedures that up to now is held only by your most experienced employees. At the same time, codifying processes is also an opportunity to make them more efficient.
Increase process throughput by up to 4x with existing employee headcount.
Relieve employees of mundane tasks and free them up for more rewarding work that’s more valuable for your bank.
Keep clients happy by getting deals done more quickly and accurately.
While analyst firms including the Everest Group use the term intelligent document processing, the idea of applying artificial intelligence technologies to process automation can go by various other names. The term “intelligent automation” is also common and Indico long ago coined the term “intelligent process automation” (although we’re flexible).
Gartner lumps IDP with other automation terms under the umbrella of “hyperautomation.” Its definition reads: “Hyperautomation deals with the application of advanced technologies, including artificial intelligence (AI) and machine learning (ML), to increasingly automate processes and augment humans. Hyperautomation extends across a range of tools that can be automated, but also refers to the sophistication of the automation (i.e., discover, analyze, design, automate, measure, monitor, reassess.)”
Choose whatever term you like, but when it comes to selecting an IDP platform, choose wisely.