So says Rob Stone at SS&C Technologies, in this FTF News Q&A. Stone provides details about SS&C’s digital workers, an AI Gateway and A.I. tools guiding SS&C’s products and services.
Rob Stone
(Applying artificial intelligence (A.I.) to securities operations is “definitely a force multiplier,” says Rob Stone, senior vice president and general manager of intelligent automation and analytics at SS&C Technologies, in this FTF News Q&A. For the past two years, Stone has been getting to know all that SS&C Blue Prism can do for A.I.-powered automation, particularly services and offerings from SS&C. In this chat, Stone provides details about SS&C’s 3,000 discreet digital workers, an AI Gateway administrator interface for large language models, and how A.I. tools can sit atop SS&C’s products and services. Stone oversees SS&C’s Blue Prism, Algorithmics, Distribution, and Regulatory Solutions businesses.)
Q: There is a hype train around artificial intelligence that’s annoying, but at the same time, you can’t ignore it. Not within financial services in 2025.
A: No, A.I. is definitely a force multiplier. The speed at which you can take action based on the use of large language models [LLMs] and other A.I. tools is just eye-watering. We’re excited about it. We’re testing heavily internally, we’re utilizing these tools in a secure manner, and we’re building capability that our clients are going to be able to take advantage of.
Q: Are there particular SS&C offerings that are using A.I. and are getting traction?
A: In March 2022, SS&C completed the acquisition of Blue Prism. For the last two-and-a-half years, we’ve dedicated a ton of effort and resources to the internal deployment of our automation suite within SS&C.
Today, we have 3,000 discreet digital workers that are working alongside and on behalf of our own services professionals to automate non-thinking tasks. It’s the largest and fastest deployment of Blue Prism technology in the history of Blue Prism. The technology is used in hedge fund and private equity administration, loan processing, and mutual fund transfer agency in regulatory compliance — all of the major functions that SS&C handles on behalf of our clients.
We’ve got scaled deployment of a digital workforce, and we’ve got hundreds and hundreds of people now that are experts in that automation deployment. … It gives us an incredible development feedback loop of people that are using this commercial product themselves internally at SS&C, giving us a bird’s eye view in terms of what more they’ll need to do their job more effectively. Almost resoundingly over the course of the last 18 months plus, it’s been: ‘Can I have access to Large Language Models in order to become a force multiplier of my ability to automate even more and also automate a higher level of complexity type tasks?’
The Blue Prism unit and the center of excellence within SS&C that supports the deployment of Blue Prism across our services businesses started looking at providing that access to large language models, but quickly came to a realization that there was an incredible amount of risk in providing unfettered access to A.I. tools. So we started to build a security framework around that access and guardrails to allow for the responsible deployment of A.I. across SS&C’s vast workforce. We built a system called AI Gateway that provides an administrator interface on top of various LLMs to allow for the redaction ability of sensitive information from any information that was going to go into an LLM.
Q: And this is strictly internal, or is this also offered to your clients?
A: It started as an internal platform with the eye to be commercially available. We’re now at the point of making AI Gateway commercially available to the market.
Q: When is that going to happen?
A: That’s in progress. We’re in the alpha/beta stage of the process. We’ve got it up and running. We’ve got hundreds of internal users in production using AI Gateway to access LLMs to automate processes using A.I.
Q: And when you say digital workers, is that a nice term for robots?
A: It is a nice term for robots, and it’s increasingly a shift to A.I. agents. So, it’s becoming a mix. It was originally robots, traditional RPA [robotic process automation]. But as we started to allow for A.I. to be integrated into the automation suite of SS&C’s processes, it’s quickly becoming more and more A.I . agents that are operating within SS&C. And all of those A.I. agents that we’ve got internally — and we’ve got dozens of them, many of which are at that precipice of being ready to be sold as commercial software as A.I. agents that people can buy from SS&C — they’re all running on top of or within the AI Gateway framework.
Q: There has been a lot of innovation within SS&C, and there have been a lot of acquisitions. Will A.I.-based technologies help you unify companies and their offerings that were developed separately and now are under one umbrella?
A: The traditional RPA has done that in a big way for us. Through acquisitions, SS&C tries to add capability, and one example of where Blue Prism’s added value is in the SS&C Algorithmics space, where we’ve got large banking institutions that have deployed digital workers that are configured for end-user risk management workflows.
The combination of Blue Prism with vertical products for specific functions inside institutional financial services firms has already been driving results in terms of unifying SS&C’s products and services. Whether you’re talking about the big global multinational financial institutions like banks or insurance companies, or institutional asset managers, many use a multitude of SS&C’s products and services. So, really, what you have is a dataset that is very useful for reporting and compliance, and analytics that needs to be accessed, normalized, and enriched, and then brought together in order to provide as much value to the top of the house as possible.
When you think about what A.I. does in terms of fast data access, we think generative and agentic A.I. tools sitting on top of SS&C’s products and services inside an individual customer unlocks the value of data in a big way. That’s part of SS&C’s strategy going forward.
Q: So, stepping back a bit, what does AI Gateway do? How does it control the LLMs? And does it also oversee other aspects?
A: Well, it’s really about where the LLMs are run. So we’ve got GPUs [graphics processing units] running in SS&C that allow us to, within the four walls of SS&C, access large language models where that data SS&C is feeding to those LLMs doesn’t leave SS&C’s premises.
AI Gateway gives users the ability to upload information to extract analysis based on that information without making sensitive information leave those four walls. You’ve got an administrator-type user interface. Administrators can put in rules to redact phone numbers, Social Security numbers, security identifiers, and various things that you wouldn’t want going into a black box assistant. Then we’ve been able to build A.I. agents on top of that. We’ve got invoice reconciliation agents and private equity partner capital statement agents, and credit agreement reconciliation agents that we’ve built on top of the expertise and technology that SS&C uses to service its customers. We’re now commercializing them for external use.
Q: What are your clients either looking to implement or are implementing as far as enterprise-wide automation within the trading realm, securities operations, particularly the middle office, and the back office? What are their concerns?
A: I think the concept of enterprise automation is where you’ve got workflows that touch different departments, systems, and data sets across different functions of a big institutional customer.
What you need is the ability to orchestrate how those automations play together. The ability to cut across all of those various departments, functions, and data sets to achieve an outcome really relies on enterprise orchestration. A lot of our customers are asking us for more and more orchestration tools that let them create automations that are truly enterprise and not just desktop.
The automation space is bifurcated between desktop and enterprise. In a desktop setup, an individual has access to an automation, whether it’s a bot or an A.I., to automate workflows that they do every day. Whether that’s organizing an email inbox or whether that’s quickly downloading documents and extracting data to upload into another system, those desktop automations are important, but they’re not really scalable. They’re not really enterprise because they’re focused on what an individual is doing on a day-to-day basis.
When it comes to automating regulatory reports to the SEC, you’re looking at access to information from internal datasets and external datasets. You’ve got to go to different brokerage houses, you’ve got to go to different custodians, perhaps. You’ve got to have access to the portfolio system, to the partnership accounting system. You’ve got to organize data from risk systems in order to answer specific questions on, say, Form PF.
In order to do that in an enterprise way, you need a toolset that allows for various levels of access to pull together, enrich, upload, validate that data in a way that doesn’t take time away from people to do more high-value work. I think what people are looking for from an enterprise automation solution is really to free up their people to do client-facing work, analytics, high-value work, and allow for data access protocols, data reconciliation, and data uploads to be handled by bots.
When you’re talking about enterprise workflows that include financial statement prep or tax reporting, or regulatory reporting, there’s so much information that goes into those high-impact reports that desktop automation really isn’t able to achieve that enterprise goal or enterprise outcome.
Q: How will A.I. impact the back-office operations of clearing and settlement, and reconciliation?
Q: I think it’s going to affect all aspects of middle and back-office processing. From trade settlement to tri-party reconciliation to P&L generation to collateral management, we’re seeing more efficiency within our own operation. And we operate on behalf of our customers, just like our customers operate.
SS&C has nearly $3 trillion worth of alternative assets where we’re providing full middle-to-back operations to accounting services. We’ve got thousands of people receiving trade files, providing trade communications, settling and confirming trades, reconciling those to the street and to internal party systems. We’ve got A.I. and robotics deployed in production, assisting with the more efficient and scalable way to process all of those activities. We’re updating payment systems following counterparty confirmation emails, which are enabling efficient pre-settlement processing. We’ve got margin call agents that are processing daily margin call reports.
We’re extracting SMI [Stochastic Momentum Index] summaries and preparing end-of-day uploads. We’re streamlining that broker management process. We’ve got OTC [over-the-counter] confirmation agents validating, extracting and validating key trade data to support downstream operations through to settlement, legal and compliance—things like reviewing contracts to extract information from contracts like CPI [Consumer Price Index] clauses, analyzing contracts for key legal and risk terms, and providing summaries to the finance and compliance functions.
We’re seeing efficiencies through the use of A.I. in our own operations, which our clients are obviously able to take advantage of through a relationship with SS&C. And we’ve also got an eye on taking those internal use cases that we’ve built in our services business and turning those into an A.I. platform that our clients can buy directly from us.
Q: What about artificial intelligence and operations makes you a little nervous, and maybe keeps you awake at night?
A: The thing that really tempers our enthusiasm is the reliability aspect. As the A.I. hype train has grown and the promise that A.I. shows is unbelievable, the stakes at which our customers are operating are extremely high. So the room for error is very low. And when you think about not only data loss protection as being extremely important, but also the output that you’re receiving from an A.I. model, you are still worried about the risk of hallucinations. In financial services, when you’re moving millions, hundreds of millions, billions, if not trillions of dollars, you have a pretty low room for error. Just like the risk of a fat finger can cause a lot of pain for financial institutions, so can an A.I. hallucination.
Those are the types of things that keep us up at night and why we’re being pretty circumspect as to how we’re implementing A.I. on behalf of our customers and with our customers in really high-stake situations— due to the sensitivity of the data that we’re in custody of, as well as the sensitivity of the actions that are taken based on outputs that are calculated by A.I. models.
Need a Reprint?