Key Takeaways:

• How should firms approach a platform transformation, and when should they “buy vs. build”?
  Transformation must be business-led with clear executive vision for both current operations and the desired end state. Firms should build capabilities that capture their unique value proposition or institutional edge, while buying or licensing established systems of record like accounting or trading platforms.
• What key software engineering challenges and guardrails must be considered when building applications in the era of AI?
  In regulated industries, security, compliance, and data privacy must be directly engineered into applications rather than left to AI tools. While AI boosts developer productivity, traditional software engineering disciplines like change management and system controls must still be enforced. Standard interfaces like APIs or MCP servers should be used to encapsulate applications and prevent breaking downstream workflows.
• How can an organization manage and structure AI adoption safely?
  Monroe Capital deployed an enterprise AI tool using monthly 50-person cohorts with dedicated training to build personal productivity skills. For coding capabilities, they established a small “citizen developer network” operating within a controlled sandbox environment. These pilot projects are overseen by senior management to ensure alignment with business strategy before broader deployment.
• How do you filter through the constant noise of AI vendor demos and decide which products to evaluate?
  Evaluation relies on prioritizing tools that fill existing operational gaps or offline tasks rather than replacing core systems of record. He sets a much higher bar for replacing foundational systems than for adding new capabilities. Embedded AI functionality and support for protocols like MCP are now baseline expectations for modern SAS platforms.
• How do you define success and measure ROI on multi-year transformation initiatives and AI investments?
  Transformation success is measured by overall business outcomes—such as scaling assets under management without a linear growth in headcount—rather than strictly following shifting project timelines. AI ROI is tricky to quantify because costs are concrete while returns are hard to measure. ROI is realized when tools enable teams to handle significantly higher volume or free up staff for higher-value work.
• What foundational advice would you give to firms undertaking transformation or expanding their use of AI?
  Well-architected and governed data is the essential foundation for both traditional platforms and AI tools. Because AI accelerates data usage and amplifies existing data quality issues, strong data discipline is mandatory. Organizations must ensure their underlying data structures can cleanly expand to support dynamic reporting and analysis.