The Future is Now: A Temperature Check

• At a recent conference, the prevailing mood surrounding AI adoption in Investment Banking felt like a prolonged sense of anticipation.
• Productivity and top-line gains feel a bit farther removed than optimists hoped for a year ago, but there is an improved line of sight. 
• AI solutions that deliver role-specific productivity gains for higher-end knowledge work are increasingly the focus.
• Over the next 12 months, ModuleQ will continue to deploy AI tooling that aims to reshape the investment banking workflow. 

This post is the third in a series entitled The Future is Now. Read Posts 1 and 2. 

From AI Hype to Banking Reality: The Scaling Struggle 

In prior posts, we highlighted the need for banks’ AI investments to start delivering productivity and top-line gains. This has proven harder than flipping a Generative AI switch. Nevertheless, banks and their innovation leads have found several interesting paths forward. These include a focus on enterprise search enhancements, as well as targeted applications of Gen AI in siloed verticals such as customer service / live chat, software development, wealth management, and document review/reconciliation. 

And yet, there is a prolonged sense of anticipation—an expectancy. “Where is the beef”, one might ask, if they are old enough to remember that advertising campaign. The prevailing sense certainly isn’t negative. It’s more like the exciting electricity that fills the air leading up to a big race. 

That was our takeaway at this year’s LSEG Financial Markets Connect 2024 Banking Panel in London, where ModuleQ CEO and co-founder David Brunner spoke. The enthusiasm from last year has carried forward, as has the focus and commitment to AI investment. But the productivity and top-line gains feel a bit farther removed than optimists hoped for a year ago. This will require work. Here are some reasons why: 

• There are inherent Gen AI characteristics that create a steeper energy of activation, especially when applied in regulated industries where governance is a gauntlet. 
• Scaling up is hard! It’s easy to get something demonstrable in a lab, which wows viewers. But making it industry-grade is not as simple as it looks. We liken this to the old world of enterprise software deployment (servers and all) vs the newer world of enterprise SaaS (downloading apps). 
• AI deployment is more of a journey than the traditional form of software procurement. It requires learning by doing. According to a recent study by Bain, firms with more automation experience are adjusting quicker, but it is an adjustment for all. 
• Top-down solutions meant for all workers are proving challenging to implement in bespoke roles such as investment banking. This is due to the need for accuracy, specificity and domain calibration. And so, an AI assistant is only as good as the accuracy and relevance it can provide. 

The Next Twelve Months: Moving to Tailored AI Solutions 

From our interactions with bankers and partners at the conference, we are seeing a few interesting trends for the next twelve months. First, many firms are applying machine learning to existing optimization problems, with the umbrella label of AI. It’s a laudable improvement that sidesteps some of the concerns surrounding Generative AI. Other firms are taking a measured approach towards any AI adoption, focusing less on unknown productivity gains and more on needle-moving top-line gains. We tend to agree with this prudence but caution against staying on the sidelines too long. Finally, an overarching sense is that many firms are examining workflow automation but are doing so with a refined understanding of Gen AI’s current limitations, compared to 6 or 12 months ago. Many are finding that complementary approaches may prove more actionable. 

And so, AI solutions that deliver role-specific productivity gains for higher-end knowledge work are quite the focus. Paradoxically, they are rare given the inherently challenging lift. Firms like ModuleQ are trailblazing through new pathways like Unprompted AI. This requires broad education. We also believe in enterprise worker personalization, another nascent concept that will require a mind shift. Here are a few things ModuleQ will be actively doing in the next 12 months to practice what we preach: 

• Continue to proselytize “push” as a complementary modality for leveraging AI. 
• Amplify personalization as a means of routing superior information to specific bankers. In doing so, we can better understand the core persona of a knowledge worker, tailoring the information that makes them more productive. 
• Roll out exciting new product functions that expand the remit of the imaginable, accelerating adoption. 

The Future Investment Banking Workflow: Personalized, Integrated, Unprompted 

The future is bright. We are seeing glimmers of what a new investment banking workflow looks like. It is about better information connectivity, leading to better integration of systems and data sets, equipping a new-age banker. It’s about push as well as pull, where information is surfaced to a banker, based upon their knowledge graph. It’s about hyper-personalization: the application of software and data to their unique needs, with the capacity to dynamically evolve with them. All of this will ultimately drive a new era of investment banking. The future is indeed now for investment banking. We just need to push in the right direction.