Mogo / Senior Marketing Operations Manager / 2024 – Present
Agentic ops infrastructure
Leadership mandated automation with no platform and no owner. The real need was ops self-sufficiency.
The read
The directive came down: we are automating. No tool selected, no owner named, no scope defined. The ops team was executing manually on loan routing, compliance review, and email production. Leadership wanted AI involved. The team wanted to keep working.
The read: the request was not really about AI. It was about reducing the manual bottlenecks that were slowing everything down. The team needed to be able to do more without adding headcount. The tool was the vehicle, not the point.
The decision
Pick three high-friction manual processes and build agentic workflows for each. Start with loan routing, compliance review cycles, and email production. These had the clearest input-output logic and the most measurable time cost. Build the infrastructure for the team to run them, not for a central team to manage them.
The selection criteria was: where does a human spend time doing something a model can do more consistently? Not: where does AI sound impressive?
The metric I chose
Cycle time reduction per process. How long did it take before vs. after? I tracked this separately per workflow so we could see which interventions worked and which needed adjustment. Secondary metric: underwriter close rate on AI-routed loans.
The build
Built an AI loan routing agent that scored inbound applications and matched them to the right underwriter tier. Built a compliance review workflow that drafted and flagged review items, reducing the manual read cycle. Built an email production system using AI to produce first drafts from a creative brief, reducing build time by around 75%.
Each workflow was designed so the ops team owned it directly. No central AI team required. The infrastructure was the point.
up ~30%
Underwriter close rate on AI-routed loans
down ~50%
Compliance review cycle time
down ~75%
Email build time
The routing improvement was the most significant. Better loan matching meant the underwriters were spending time on applications with higher close probability. The email and compliance gains were efficiency wins, but the routing improvement directly affected revenue.
Tech
OpenAI API, loan routing and compliance review agents
n8n, workflow orchestration
Braze, email production pipeline
Confluence, workflow documentation and SOPs
Miguel N. Monzones
Vancouver, BC, Canada
