Main content

Snowflake Summit 2024 - how JP Morgan uses consistent data sources to help its clients make smart decisions

Mark Samuels Profile picture for user Mark Samuels June 6, 2024
This novel partnership between Fusion by JP Morgan and Snowflake draws on emerging technology to solve the intractable challenges of institutional investors.

jp morgan

JP Morgan is using Snowflake technology to ensure its clients have access to a trusted and consistent data source for crucial financial decisions.

Gerard Francis, firmwide product head for Data and Analytics at JP Morgan Chase and Head of Fusion by JP Morgan, says organizations that want to exploit analytics, AI and machine learning must have access to useable data in a modern technology stack. That’s where Fusion by JP Morgan comes in, allowing investors to retrieve investment data held by JP Morgan’s Custody, Fund Accounting and Middle Office services, using cloud-native channels, including the Snowflake Financial Services Data Cloud.

JP Morgan has a security services business that manages billions of dollars of assets under custody for institutional investors, such as asset investors and asset managers. The security services business runs much of its investors’ middle and back-office work and holds information on their transactions and positions. This information, when combined with other data, can be tough for clients to manage, explains Francis:

Institutional investors struggle because they have all this incoming data, either from people like us in security services, or from their market data vendors, or their internal data. All of this data is disparate and doesn't connect and talk to each other, and it's quite noisy for them to deal with.

He says clients recognize that JP Morgan holds their transaction and position data and are keen for the organization to make all information available consistently. Many of these investors are Snowflake users, so they’re eager to receive consistent data in their Snowflake instance, as well as via other database technologies and APIs, he adds:

They want the whole problem solved...What's important is that we make the data work natively in our clients’ Snowflake instance.

Francis joined JP Morgan three years ago and has been using Snowflake with Fusion to solve customers’ data-feed challenges for the past two years. Fusion also delivers consolidated data to clients’ other technology platforms, including Databricks.

Francis says working with Snowflake provides a range of benefits, not least consistency. He describes financial data as being “remarkably complicated”. Fusion has to deal with as many as 50 financial instruments, from bonds to mortgages and equity options:

The data itself is hard to manage. There are also no standard identifiers in the market that work across the board. There’s this overwhelming amount of data. And there's no easy way to figure everything out. It's a jigsaw puzzle.

Fusion has a feature called the Spine which takes different data types and uses algorithms to sort information on the customer's behalf. These disparate data feeds are consolidated into one consistent format and are available in a client’s Snowflake instance, Francis explains:

So regardless of your source, we'll make all your data look and feel the same. The data lands in whatever operational technology stack the customer has and it works. If people want, for example, deep historical data going back in time, we'll solve the problem for them.

Embracing AI to boost innovation

Francis says Fusion by JP Morgan and Snowflake have worked well together to identify and solve institutional investors’ challenges, including using Snowflake’s Cortex AI solution to help clients use natural language to query their data.

His team in Fusion created a single semantic layer across all financial data. That’s an important step because finance data is often inconsistent and similar fields are given different names. Francis says one consistent data set makes it easier for Large Language Models. His team ships a container of data with the semantic descriptions to Snowflake and clients can use Cortex to get answers to their most pressing financial questions:

People can write natural language questions and get amazing responses to hard questions. A combination of Fusion data and the semantic layer with Cortex brings that solution to users.

Francis thinks his organization’s pioneering use of AI and Cortex could be the first important step in a longer-term relationship:

We'll continue to partner with Cortex because its capabilities will grow as Snowflake makes its queries more powerful. We'll ensure that we continue to work with them as both partners grow the richness.

He says the key to success will be finding innovative ways to ingest complicated data and make it consistent and simple for clients, regardless of the use case:

I suspect there will be more capabilities in the future, whether it's data quality or other areas, where we’ll partner. When you have great analytical partners, you want to leverage their capabilities so clients can get the business results without doing the hard work.

A grey colored placeholder image