Apex Fintech Solutions has launched its Apex AI Suite, featuring one of the first agentic development kits in the clearing and custody industry.
The tool is intended to accelerate developer integration with the company’s AscendOS platform through natural language commands and, in turn, enable faster development of wealth platform prototypes, cutting development time from weeks to days.
Such innovation could help drive growth at Apex Clearing, the clearing and custody unit and wholly owned subsidiary of Apex Fintech Solutions, which custodies more than $229 billion in client assets across more than 22 million brokerage and other types of client accounts on the platform.
The developer kit is intended to operate alongside Ask Ascend, Apex’s AI assistant launched in March of 2025 and built on top of Google Cloud’s Vertex AI platform, which also forms the foundation of the Apex AI Suite.
“Apex is fundamentally changing who gets to build in financial services,” Bill Capuzzi, CEO of Apex Fintech Solutions, said in a statement. “You shouldn’t need a team of engineers to test a great idea. We’re putting the power to build wealth platforms directly into the hands of the innovators, turning vision into a working product faster than ever before,” he said.
NVIDIA, in its sixth annual State of AI in Financial Services report published earlier this month, found that while 42% of the 800 respondents from around the globe and their companies are using or assessing agentic AI, only 21% reported having already deployed AI agents.
According to Apex, the newly released toolkit will enable business leaders to prototype and validate features without engineers, accelerating development cycles for both technical and non-technical teams. Simply put, AscendOS is meant to be a cloud-native, vertically integrated platform that enables trading, custody, reconciliation and investor documentation on a single infrastructure rather than spreading them across different providers. This latest announcement, the firm hopes, will allow a wider evaluation of custodial platform development, extending the process beyond an engineer-only approach and making it a hands-on experience for product managers, executives and business stakeholders.
Asked about the potential of this rollout, industry experts and analysts expressed optimism, tempered by some concern.
“We’re encouraged to see Apex pushing agentic tooling into the clearing and custody layer,” said Haik Sahakyan, CEO and co-founder of ARQA, an end-to-end AI-native wealth management platform (ARQA won a 2025 Wealth Management Industry Award in the Technology Disruptor category).
“Building on Vertex AI for enterprise-grade security and enabling non-technical teams to prototype via natural language is a real step forward, especially for firms that are deeply concentrated on Apex and AscendOS looking to embed the user experience in their platform,” Sahakyan said.
He said that, based on his company’s experience building across custodians, the limitation is that the ADK is inherently platform-bound, meaning it is powerful when most of your assets and workflows live in a single ecosystem.
“That’s not how most RIAs actually operate—the majority are multi-custodial, while also pulling from separate CRMs, performance systems, planning tools, and document workflows—a tightly coupled ADK risks accelerating innovation within a silo rather than across the firm as a whole,” he said.
First announced in January 2024, AscendOS, Apex’s cloud-native trading, clearing and custody platform (initially referred to simply as Apex Ascend) became available in July 2024.
In addition to allowing Apex client advisory and wealth management firms to manage traditional assets, Apex later launched Apex Alts on Ascend a year ago, enabling the management of alternative investments.
According to Apex, the ADK bundles AscendOS API references, guides, code samples, software developer kits and a knowledge base into downloadable files that integrate directly into a developer’s terminal or integrated development environment. With the tools, users can generate code with natural-language commands like “implement account opening” and receive what is ostensibly production-ready code in Java, Python, TypeScript, Go, and other languages.
“When managed well, agentic AI supports growth by improving client acquisition and increasing operational efficiency while complementing human judgment,” said Uğur Hamaloglu, EY Americas Wealth & Asset Management consulting leader.
He said that firms adopting agentic AI must balance its benefits with strong governance that preserves human oversight and clearly defined fiduciary responsibilities.
“This helps reduce the risk of investors acting on AI‑generated recommendations without professional guidance,” he said.
In its 20th annual Cost of a Data Breach Report, released last year (subtitled the AI Oversight Gap), research from IBM and the Ponemon Institute illustrates how AI is already outpacing security and governance. The study, which examined 600 organizations of various sizes across 16 countries and geographic regions and 17 industries (14% of respondents were from the financial industry), found that 63% of organizations lacked AI governance policies to manage AI or prevent the proliferation of shadow AI.
Alois Pirker, a longtime financial services industry analyst, researcher and founder of Pirker Partners, said that another consideration with growth in agentic AI is the need for wealth management firms to exercise sufficient quality control before going live with agents.
“Businesses using many agents will create a big maintenance headache, and creating a change management process will be vital,” he said.
If not conceived well, he said, agentic AI will create an effect similar to that seen with the rollout of robotic process automation bots in recent years.
“Just because work can be automated by using one of them, doesn’t mean that they don’t step on each other’s toes or might operate on flawed data, despite being very efficient, it still results in flawed results,” Pirker said, reiterating the need for strong monitoring and governance.
ARQA’s Sahakyan said his company has already seen similar issues at some of the larger enterprises it has worked with, as it focuses on custodian-agnostic orchestration and unified intelligence by employing its own multi-agent platform, rather than tying agents to any single platform.
“Empowering non-developers is valuable, but without firm-wide orchestration, shared context, and oversight, it’s easy to end up with multiple agents doing similar things across departments; each with different assumptions, data sources, and compliance implications,” he said.
Sahakyan was nonetheless unequivocal about the technology’s ultimate effect on the industry.
“Agentic AI will absolutely reshape wealth management,” he said.