Apex E3: Helping customers make smarter real-time investment decisions with generative AI

Capital markets are awash with data. However, the volume and complexity of that data makes it difficult for banks and asset management companies to leverage it effectively to inform their investment decisions. With its focus on real-time big data analytics, Apex E3 helps large asset management firms and banks solve this big data problem. By combining financial services companies' stores of underexploited data with real-time market analysis, Apex E3 provides asset managers with instant insights, such as predictive modeling and forecasting, to allow them to make better investment decisions.

Such complex real-time data analysis requires significant processing power. However, when it was previously running on a third-party cloud provider, Apex E3 experienced reliability issues which impacted the services it could offer its customers.

"When we use a server in the cloud, we need it running at 100%, all the time," says Usman Khan, CEO and Founder, Apex E3. "Instability was a challenge for us. There was a period when we were doing a release over a weekend and we lost all access to the servers, delaying the release. That was a key factor in our decision to find another cloud provider."

The company's need to change providers was motivated by other factors too. Hidden costs made it difficult to optimize its cloud expenditure, while the need to use encrypted keys to log in to its virtual machines made managing access permissions a time-consuming process. When Khan learned that it wasn't necessary to use keys to access Google Cloud terminals, without jeopardizing security, he realized it would make identity and access management far more straightforward for his team.

"With Google Cloud you can securely access the console directly from the webpage, making that whole headache go away," says Khan. "From an operational perspective, that's a massive win for us. And then it's one click to remove someone when they leave, versus having to regenerate keys for everyone. With Google Cloud, it's so much easier."

After migrating its workloads to Compute Engine in less than a week, with no downtime or impact on its customers, Apex E3 immediately noticed an improvement in the performance and reliability of its servers. The company has experienced 99.9% availability to date, while benefiting from more CPU capacity than it had with its previous cloud provider, meaning it can do more computation on less hardware. As a result, Apex E3 is able to deliver new releases to customers more quickly, taking advantage of existing server capability, rather than having to spend time and resources adding new servers.

The ability to spin up new virtual machines (VMs) quickly is also saving Apex E3 time. Previously, adding a new machine could take a couple of hours, whereas now it can be done in two clicks with Google Cloud, taking just a few minutes, allowing the team to deliver releases more quickly to customers.

The migration to Google Cloud has helped Apex E3 reduce its cloud spend, too, as Khan describes. "Expenditure with Google Cloud is very transparent, with no hidden costs, and if you want to optimize further, it's far more straightforward than with our previous provider. Compute Engine is also 20% to 30% cheaper overall than running similar VMs on our previous infrastructure."

With its workloads now running on Google Cloud, the Apex E3 team has found it easy to experiment with AI models in the Google Cloud Model Garden on Vertex AI. As Khan explains, being able to access and experiment with these services quickly and easily has enabled the Apex E3 team to innovate faster. "If we're trying to solve a specific use case for a customer, we would previously have had to set those models up ourselves. Now, because we can access Google Cloud models in Model Garden, we can do weeks' worth of work in an hour. That has a huge effect on our release and research and development timelines."

That ability to experiment quickly has led Apex E3 to use Google Cloud generative AI models to help its customers gain faster, smarter insights from their data. In particular, Apex E3 has used PaLM, Codey, and Gemini models in its document analysis tool, enabling customers to gain insights from large documents, such as research publications on credit spreads, without having to read the whole text. Powered by Google Cloud large language models, the tool allows customers to use natural language to ask very specific questions and get clear answers immediately.

"The new Gemini 1.5 Pro is an amazing model," says Khan. "It summarizes all the pertinent points in a much more concise way and you don't have to spend a lot of time trying to engineer the prompt to get the right answer."

Apex E3 is also using Gemini models to build more complex generative AI tools capable of analyzing documents containing both images and text at the same time, for example an article containing written market research alongside graphs. With Gemini generative AI models, Apex E3 enables its customers to ask questions of the document and gain powerful, instant insights based on both the text and the graphs. As a result, Apex E3 customers no longer have to wade through lengthy documents to gain the insights they need to make the right investment decisions. Apex E3's analysts are saving time, too, with the help of Gemini Pro Vision, which, says Khan, "is very powerful at analyzing charts and extracting tables."

Apex E3 is continuing to add new solutions to its pipeline, with plans to use Document AI to allow its customers to parse thousands or even millions of long documents to gain insights to inform their investment decisions. At the same time, it is considering integrating AlloyDB and BigQuery to enhance the range of data storage, access, and analysis services it can provide.

For Khan, this ability to quickly integrate new solutions is key to Apex E3's future success. "The fact that Google Cloud makes it really easy for us to interact with its services helps us to deliver features more quickly," he says. "That gives us an edge over other players and helps our customers make faster decisions. Ultimately that's a huge benefit for our customers."