Keynote interview

How to drive transformative artificial intelligence in fund services

Demetry Zilberg, Chief Technology Officer was interviewed in PDI’s Tech, AI & Fund Services edition​ about the Alter Domus technology journey, as well as artificial intelligence and its impact on private credit. He highlights the importance of managing the gap between hype and practical business applications, prioritizing opportunities and mitigating risks.

technology data on boardroom screen plus people in meeting


Q You have recently joined Alter Domus as chief technology officer. What background do you bring to the role, and how is that perspective useful for Alter Domus and its clients?

Most recently, I was a chief technology officer at Wells Fargo bank. Before that I was the CTO of the financial data and software company FactSet, which was my final role there over a 20-year period. In these positions I’ve been in­volved with a vast array of initiatives, from developing mobile capabilities and addressing challenges such as iden­tity protection and fraud prevention to developing and executing artificial in­telligence strategies and implementing generative AI.

Alter Domus is my first experience in a private equity-owned company. The mandate of the senior team here is clear: how do we use technology to make our services more efficient, more intuitive, and more impactful for our clients?

At Alter Domus, my team owns all the technology and app development efforts for the business, as well as the data analytics and engineering and some of the product strategy from a technology perspective. The goal is to help continue to scale an already rap­idly growing business and drive sever­al transformative initiatives, with one area of focus being the data and ana­lytics business, where we see a lot of potential.

Q What is your perspective on AI and its adoption in private markets?

AI has been fundamental to many com­panies and products for a number of years. Companies such as Tesla have been working on self-driving vehicles for more than 10 years, and those are in­herently AI-enabled. So, AI is not new.

What has changed is that last year saw probably one of the most suc­cessful marketing campaigns ever launched when Open AI unveiled its initial version of ChatGPT. We all started talking about generative AI and what ChatGPT would mean for us. Adoption of that tool went from zero to 100 million active users in just two months, making it the fastest growing consumer application in history. That showed the art of the possible with generative AI, and immediately drove that conversation in boardrooms.

Right now, what is important for business is to thoughtfully manage the gap between AI hype and practical business applications. Businesses need to prioritize opportunities in order of obtainability and then manage commu­nications with stakeholders to create meaningful outcomes. Too often, com­panies become focused on what’s new, rather than what’s best, or more to the point in Alter Domus’s case, what’s best for our customers.

Q What steps are service providers like Alter Domus taking in AI?

We believe that before any concrete actions are taken in AI development, governance issues have to be addressed. Protecting both our clients’ data and business at large as well as our own is of the utmost importance.

It’s best practice to have a formal and mature intake process for AI ap­plications and a multidisciplinary panel that considers and prioritizes oppor­tunities for the business, as well as any implementation guardrails and cy­ber-security considerations. That panel should include representatives of tech­nology, business, regulatory compliance and HR. Putting that disciplined collec­tive effort into mitigating risks like data leaks or data breaches is essential.

The other element of this is the technology itself. The way a company like Alter Domus approaches that is with a platform construct. Instead of focusing on underlying infrastructure (such as GPUs), we chose to partner with cloud providers that offer robust ‘out of the box’ solutions that enable us to experiment, validate, and deploy AI solutions with little friction. Leverag­ing our partners that have robust end-to-end services that aggregate versus trained or partially trained algorithms is a smarter method for us.

For example, we have a relationship set up with AWS using some of their services, such as SageMaker and Bed­rock, which deliver pre-packaged plat­forms of large language models. That is central to our strategy because we get immediate access to enhanced technol­ogy and capabilities.

One of the ways we approach AI gov­ernance is to consider use cases in three buckets: ready-made solutions we can buy (such as co-pilots that can be easily integrated into our workspace produc­tivity tools); capabilities and tools, built by Alter Domus, that enhance the effi­cacy of our employees’ service delivery to clients; and AI capabilities embedded into products, delivering sophisticated analytical capabilities and meaningful insights to our clients.

Q Why is private credit a particular area of focus, and what advances are being made there?

We feel the private credit space really lends itself to the last two of those three categories. There are tens of millions of unstructured documents and data sources in private credit, with many in PDF, Excel or even fax format.

We have a product called Digitize that processes 30 million documents, extracting, classifying, and incorporat­ing content into client-facing products. Then it becomes easier to use genera­tive AI and the analysis of that data re­quires less human input, leading to bet­ter outcomes, faster turnaround times and cost savings for clients.

Q Is AI replacing humans in the use cases you see?

I think for the foreseeable future, AI is about creating substantially better tools for humans to use, but we still need hu­mans. Instead of using manual tools to deliver outputs, you can speed up processes and get more reliable results using AI, and that is more rewarding work for the humans involved.

For now, we still need humans to oversee those processes, to check that AI is working properly, which means there is an upskilling opportunity and AI is simply taking away the more mundane elements of tasks. In the longer-term, AI will free our teams to focus on more strategic areas of work.

Q What do you expect to be the most exciting developments in this area in the near future?

We are focused on building out our data and analytics business as a com­prehensive data platform making use of AI capabilities. That will take data from various siloes and make it available in a very secure and compliant way for our clients. We plan to employ AI to enable the data platform to deliver actionable insights and new analytical capabilities for our clients.

The key element in all of this is that you have to ‘feed the machine’: AI needs data to really learn and operate in the most effective way, and we sit on a treasure trove of data. We make cer­tain to prioritize the use cases that will have a beneficial impact on our clients. We want to give them more visibility into their funds than they have ever had before, allowing them to make faster and better formed decisions.


architecture bridge traffic
NewsJune 17, 2024

Alter Domus rises again in PwC’s 2024 Observatory for Management Companies Barometer

architecture Manhattan building
AnalysisJune 6, 2024

The real state of real estate: deal volume, fundraising, and usage patterns 

Location in London
EventsJune 25, 2024

AREF Conference 2024