By Karthi Sepulohniam (Subramaniam), Head of Asia Pacific & Australia, Partners for Growth
Partners for Growth AI Series – Episode 6
Featuring Rene Shalita, Director of Portfolio Management, Partners for Growth
The best AI lessons come from the people shipping it.
For the first two episodes of this series, we spoke with PFG’s portfolio companies about how they’re deploying AI. For this one, we turned the lens on ourselves.
I sat down with Rene Shalita, PFG’s Director of Portfolio Management, to talk about how AI is changing the way we originate, underwrite, monitor, and manage a private credit portfolio. Rene’s role spans two distinct functions: portfolio monitoring and risk management on one side, and operations, enablement, and technology on the other. He manages PFG’s internal tech stack, the proprietary PFG app that consolidates data across the firm, and the tools that support our deal, finance, and investor relations teams.
The conversation made it clear that AI hasn’t reduced the amount of work for PFG; it’s expanded what’s possible. People are working just as hard, but producing significantly more.
Claude in a Day
Claude’s adoption at PFG was entirely organic. Rene set up a team license for a handful of curious people. Within a day, nearly the entire firm requested access. It became our central AI tool almost overnight.
That speed of adoption says something about the product, but also about the culture. PFG runs a dedicated AI Slack channel where team members share tools, use cases, and recommendations. Team calls include a standing segment dedicated to AI awareness. PFG requires a formal cybersecurity review before approving any new tool for broader use, a necessary discipline when handling sensitive financial data.
AI Across the Deal Lifecycle
AI touches almost every stage of a deal’s progression through PFG.
At the front end, during origination and underwriting, the team uses AI tools to accelerate research on early-stage companies, extract information from financial reporting packs, and generate initial drafts of investment memos. The team’s judgment and expertise still drive every decision, but it has dramatically reduced the time spent on manual data collection.
Once a deal closes and a company enters the portfolio, AI plays a different role. PFG has built a central database connected to Claude that allows team members across functions to query data, generate reports, and build visualizations on demand. Instead of waiting for an analyst to pull a report or a developer to build a dashboard, anyone on the team can ask a question and get an answer directly. That’s removed a bottleneck that existed for years.
One of the more interesting efficiency gains has been in product development. Previously, when someone on the team had an idea for a new internal tool, it would take multiple rounds of back-and-forth with the development team just to articulate the requirements clearly. Now, that person can use AI to build a working prototype using live data and hand it to the dev team as a fully formed brief. Iteration cycles have collapsed, and several tools we use daily at PFG have emerged from this organic build-test-deploy cycle (built by team members across the firm: investing, tech, and portfolio management).
Fraud Monitoring at Scale
The most compelling use case Rene described sits inside GFS, a sister entity to PFG that was built to fundamentally change how we monitor asset-backed deals.
The old approach was familiar to anyone in lending: borrowers submit monthly data tapes in Excel, and the team analyses them in spreadsheets. GFS replaced that with a real-time data infrastructure that connects directly to borrower systems and bank accounts, providing continuous visibility into portfolio performance.
AI has been layered on top of this infrastructure in three areas: facility monitoring for borrowing base and covenant compliance; credit risk monitoring for delinquency and collection rates; and fraud monitoring, where things get genuinely interesting.
Asset-backed financing (ABF) carries an inherent fraud risk: double-pledged assets, fabricated receivables, false documentation. Historically, verifying the supporting documents at scale was impractical. Borrowers submit high volumes of documentation in varying formats and languages, including not only properly digitized files but also pictures of physical documents that traditional text-scraping tools cannot reliably extract information from.
AI now automates extraction, translation, and verification, populating the data directly into monitoring dashboards. What would have been impossible to do manually is now running continuously in the background. Rene noted that building these dashboards alone would have taken months without AI. The team shipped them in a fraction of that time.
When AI Got It Wrong
Rene was candid about a lesson learned early in PFG’s AI journey.
An early version of the company research tool included AI-generated recommendations on whether to pursue a prospective deal. The intention was to accelerate screening. The effect was the opposite – it introduced bias. The AI was forming views based on limited, surface-level public information, and those recommendations were coloring the deal team’s judgment before they’d done their own work.
PFG pulled back. The team deliberately used AI as a decision-support tool rather than a decision-making one. The competitive advantage in private credit lies in seeing value in situations that aren’t obvious, the kind of nuanced, creative judgment AI, operating within the constraints of what it’s been given, can’t replicate. AI accelerates the process of getting to a decision. It shouldn’t make the decision for you.
Outside this project, we’ve also noticed that AI recommendations and decision-making can be inconsistent across models and depend largely on context. Because the output is not deterministic, it can be too harsh in some cases and not critical enough in others.
Turning Internal Tools Into Borrower Value
Perhaps the most surprising outcome of PFG’s AI investment has been what happened with the GFS dashboards.
They were built for internal risk management, providing PFG with a way to keep closer tabs on asset-backed deals. But the team realized the dashboards were so useful that they could be opened up to borrowers themselves.
That’s a meaningful shift. Instead of being a lender that requests reporting, PFG becomes a partner that provides analytics. Borrowers get data-driven insights into their portfolio performance, visibility they often don’t have. It changes the relationship dynamic entirely, and this differentiation wouldn’t have been possible without the AI capabilities underpinning the platform.
What’s Next
Over the next twelve months, Rene’s priorities include building AI-powered lead-generation workflows to identify potential borrowers and investors, creating an entirely new sourcing channel for the firm. Beyond that, the focus is on continuing to automate the manual processes and spreadsheet-based workflows that persist across any organization with twenty years of operational history, and developing modeling capabilities that leverage PFG’s accumulated data to generate forward-looking views on fund performance.
The broader theme is clear: AI at PFG isn’t a side project. It’s becoming the infrastructure that the business runs on.
This is Episode 6 of AI: Where Innovation Meets Execution, a series where we profile how technology companies and investors across our ecosystem are deploying AI in practical, non-obvious ways. If you’re building something interesting with AI and want to be featured, reach out to us at karthi@pfgrowth.com.
The views expressed are my own and do not necessarily reflect those of my employer.
This content is for informational purposes only and does not constitute an offer to sell or a solicitation of an offer to buy any securities. Any such offer will be made only to qualified investors through confidential offering documents. All investments involve risk, including the possible loss of principal. Past performance is not indicative of future results.
