A model for how agentic AI will change financial services
One way to think of AI agents is that it is like having an unlimited number of highly-caliber junior staff. They can reason, research, and take actions for you.
Using that as a mental model, AI agents will:
Trading – reduce market making costs.
Asset management – act as analysts and asset allocators.
Investment banking – do financial analysis.
Retail – develop more granular products.
Wealth management – inform clients.
Insurance – improve underwriting.
Operations – reduce operating costs.
Technology – increase quantity of custom technology.
These are guesses – but give us a useful “future world state” that we can imagine to think about what this means for your business.
Some implications across markets – better liquidity, better capital markets, more global markets, international competition between jurisdictions, lower unit margins for existing businesses, product innovation, and new business models.
This means that almost every financial services function will look different in five years because of agentic AI.
There are first mover gains for businesses that can adjust quickly – immediate financial, winning customers to new product types before everyone is competing, and in gathering data. This last gain will be large for many businesses – as proprietary data is a key input for AI agents.
Hiring the right investment banks for your capital markets deal
Picking the right investment bank is important. The right decision is often the difference between a successful deal and a failed deal.
The most important thing to look for is the investment bank’s access to the right investors for your deal. You also want to look for good the bankers/structurers – but the investor access part is key.
Investor access
The investment bank’s job can be look at in two parts: 1) the work to prepare the deal (how it is structured, working with lawyers to prepare the documents, etc.); and 2) selling that deal to investors.
If the bank doesn’t have the right access to the right investors – however good it is at preparing the deal, it won’t be able to sell it.
One difficulty here is that every investment bank will say it has excellent investor access. So picking the right bank needs you to work out who your ideal investors are, and then working out which of the banks you could hire have the best access to them.
Picking your ideal investors can be done by looking at similar deals to the one you are looking to do. And then finding who invested in those deals. There are databases that will provide you with investor lists for existing deals.
You then want to see which banks have excellent relationships with those investors. You can do that by contacting some of these investors if you know them, or can get introduced to them. You can then ask them which banks they like working with on your type of deal. You need to speak with the right person within the investor – the person who makes the decision for investing in your particular type of deal.
An alternative is to see which banks have worked on similar deals to yours that have been successful. This is easier for more common, larger deal types – where if the bank has worked on many of these over the last year, it is likely that they have good investor access for that deal type. You want to be careful to pick similar deals based on all the factors that matter – deal size, industry, company characteristics, etc.. For more unique deals, this can be very difficult or impossible.
Good bankers
You want to find bankers that are excellent and trustworthy.
Your bankers or structurers need to put together your deal – helping you decide on how to structure the deal, how to decide on key terms including pricing the deal – balancing what is best for investors and what is best for the company, the will help hire all the other parties in the deal like lawyers, and help put together all the materials for the deal – like legal documents and marketing material to present your deal to investors.
Excellent and motivated bankers will often be able to find options for you that other bankers can’t. Unskilled bankers will often make decisions that can result in the deal failing – or ending up being much more expensive for you than needed.
Bankers also face conflicts of interest. One big one is that you are their client, but the investors buying the deal are also their client. And unless you are a very frequent and very large issuer, the investors will often be a more important client for the investment bank than you will be. As the bank will do deals with many of these investors every week or month – and needs access to these investors to be able to do future capital markets deals. Making sure that you get good terms needs you to have bankers that you strongly trust, or your ability to know what makes sense yourself so that you are not solely relying on the bankers for your advice.
Picking skilled and trustworthy bankers is similar to finding investment banks that have strong access to your target investors. You can speak with similar firms to yours to see their experience with their bankers, and you can look at data on similar deals to yours and benchmark banks’ performance (based on the pricing they achieved) compared to each other.
AI strategy for mid-sized banks
AI is likely to change the banking industry at least as much as the creation of the internet changed the banking industry.
The problem is that we are still relatively early – new AI technologies and new solutions based on these technologies are being released every month.
So the question is what should a mid-sized bank CEO do now – if they should even do anything rather than waiting.
From working through this with multiple banks, the answer we get to is that banks should get themselves into a ready state. By getting into a ready state, banks are able to make sure they don’t fall behind their competitors – without spending a lot that could turn out to be wasted spending as new technologies develop.
To get AI ready, banks can:
Identify use cases that will show value.
Build data lakes to capture the right data.
Build and test different AI models on the data.
Upskill the business to scale AI use – by hiring people and training business teams.
Use cases could include marketing, underwriting, identifying mistakes, fraud detection, and identifying customer upsell opportunities.
Through this process, the bank gets in-line or ahead of its competitors – without risking wasted capex on solutions that will be obsolete in six months because better technology comes along, and without risking becoming locked in to technology solutions that become obsolete.
The bank instead ends up with its data in the right formats and places for any AI system to use, and it has run at least one proof-of-concept trial through which its senior and operating teams become comfortable with AI.
