About Fiduciary Trust Company
As a leading wealth management firm, operational excellence is central to delivering high-quality client service across generations.
Founded nearly a century ago, Fiduciary Trust Company was established by families, for families — with a singular focus on growing and protecting wealth across generations. The firm works with individuals, families, and foundations to build personalized investment portfolios and estate plans. Fiduciary's clients carry complex, multi-generational financial needs, and the operational infrastructure that supports them must be reliable, precise, and fast.
The 4 Layers of Process Intelligence Framework
When Fiduciary Trust began its process intelligence journey in January 2025, the team did not think of it as a reporting project. They defined a framework — four interconnected layers — that would take them from basic visibility all the way to compounding institutional knowledge.
Layer 1 — Visibility — Process Mining
QPR ProcessAnalyzer acts as the X-ray of the organization, reconstructing actual workflows directly from system event logs and revealing how processes truly behave rather than how they were designed.
What Process Mining Revealed in Wealth Management Operations
With three processes in QPR, patterns that had previously been invisible became immediately apparent.
Rejection and resubmission rates. A dashboard comparing all three processes showed, at a glance, which workflows carried the highest rejection rates and where cases were being resubmitted. In the account maintenance process, a significant share of cases — over a quarter — were caught in a self-loop between rejection and resubmission. That single insight prompted a direct line of inquiry: what information was missing? What documentation gaps were causing rework? And how quickly could the process be redesigned to address them?
Bottlenecks in cash processing. In the cash transaction workflow, completion rates on same-day items were high — expected, given the volume of wires and ACH transactions. But beneath that headline figure, approval bottlenecks were visible. The data showed where cases were waiting, and who or what was holding them. The team could immediately ask: is this a staffing issue? An automation opportunity? A policy threshold that should be adjusted?
Highest-impact outliers. A priority ranking showed which transaction types were carrying the greatest total lead time — surfacing exactly where optimization effort would generate the most return.
None of these patterns had been visible through standard reporting. They existed in the event log, waiting to be reconstructed.
How AI Generated Training Documentation from Process Data
The most striking example of process intelligence in action at Fiduciary Trust came from an unexpected direction: training documentation.
Using MCP and AI, the team queried QPR for the complete account opening process. They then asked the AI to generate a step-by-step training guide — not based on designed workflows or policy documents, but based entirely on actual process behavior extracted from the event log.
The result was a 14-step interactive training guide, complete with watchpoints highlighting where most errors and delays occur in practice. The guide included edge cases that no manually authored documentation would have captured, because the data revealed them directly from how the process actually runs. A short quiz followed at the end. Total production time: approximately 15 minutes.
What this replaced was significant. Traditional training documentation requires hours of interviews with subject matter experts, manual writing, and ongoing maintenance as processes evolve. The AI-generated guide was grounded in fact, not recollection — and it could be regenerated at any time to reflect the current state of the process.
