Enterprises need to constantly look for ways to improve and expand what they offer to the marketplace. For example, Sameena Shah, managing director of AI research at JPMorgan Chase, says the company’s bankers have been looking for new ways to study early-stage startups looking to raise capital. The challenge was, she says, “finding good prospects in a domain that is fundamentally very opaque and has a lot of variability.”
The solution for JPMorgan Chase was a new digital platform, built off an algorithm that continually seeks out data, and learns to find prospects by triaging its data into standardized representations to describe startups and likely investors. For users, the platform also offers the context of its output, to help them understand the recommendations. “Many bankers told us that they had not known about some of the contexts or data points. That’s the power of machines,” Shah says.
Embedding ESG goals in strategy
Forward-thinking financial services can also help investors that are looking beyond just the enterprise’s bottom line. Dubourg says new investments draw on a growing pool of external data to move into new investing contexts. “We’re moving from a world of unconstrained economics to a world with physical, environmental limits,” Dubourg says. Doing so, he says, means internalizing novel external data; expanding from traditional financial analysis to a model increasingly defined by nonfinancial factors such as climate change and environmental, social, and governance (ESG) goals. Given the breadth of potentially relevant data in these cases, even specialist investors and companies are unlikely to have access to all the knowledge necessary to make fully informed decisions.
JPMorgan Chase’s own solution, ESG Discovery, draws single-source ESG data from relevant businesses and sectors, providing thematic deep-dives and company-specific views. Dubourg says the platform makes sure investors have “every relevant piece of ESG information accessible in one, single spot.”
Developing innovative employees
Innovation is meant to improve how companies work, which does not necessarily involve new technologies or devices: sometimes it is a matter of rethinking processes. For this, talent is essential. An expansive approach to talent can give companies richer choices to support their work. Gill Haus, CIO of consumer and community banking at JPMorgan Chase, says developing the technology at the center of the firm is not just about finding a group of brilliant individuals, it’s about organizing around products and customers. “What really makes a technology organization,” Haus says, “is the way you hire teams and the way you coach them.”
One way JPMorgan Chase nurtures innovation is its Tech for Social Good program, focused on engaging community members, especially students and nonprofit workers. This community-based initiative is focused on developing new thinking from inside and outside the company. It has three main goals: innovate for the social sector, build the workforce of the future, and develop skills within the company. “What’s so exciting here is we have so many complex problems to solve, so many incredible people that are looking for assistance, that you just have an environment where people can grow their careers really quickly,” says Haus.
Deploying emerging technologies
Driving innovation at JPMorgan Chase focuses on finding ways to improve how cutting-edge tools are applied, such as AI and ML. To ensure responsible AI, for example, the company’s ML designs go beyond standard software development controls, or even focusing on explainability, responsibility, and training, as most companies do, says David Castillo, managing director and product line general manager for AI-ML at JPMorgan Chase. This “fairly unique” process ensures responsible AI is in place at a higher level, so that even lines of business at different maturity levels for AI and ML operate at the same standard as any other, he says.
“We’re addressing the entire machine learning development life cycle,” Castillo says. Instead of restricting innovation, this approach “creates a very interesting, streamlined opportunity for machine learning from end-to-end. We’re being responsible across the entire spectrum,” he says. “We want to be able to make sure that that every piece of data that’s being used for model training has lineage that we can trace back to its origin,” he says. It’s important that new iterations of a model feature carry forward its lineage, he says. “We’ve scrubbed that data for personally identifying information [PII], we’ve taken out proxies to PII, we’ve identified all of these landmines.”