Leslie Madrid
Innovative Instruction
.png)
AI + LD Samples
Uniform Residential Loan Application Course
Click on the image to enter the course.
Challenge
Loan officers and real estate agents who assist home buyers need to know not just what the fields in the URLA form mean, but also how to complete them accurately and guide clients through the process. Feedback from recent home buyers showed that many had trouble completing the form and were frustrated that their loan officers weren't better able to help.
I was given two sources of material to use to design and develop the course: the form and an instruction document.
My Process
I designed a structured TRACI prompt directing Claude to analyze the form and instructions, identify the most critical sections, flag commonly misunderstood fields, and produce a draft course outline with Bloom's Apply-level objectives.
Rather than accepting the AI output at face value, I reviewed it against the source documents, corrected objective language that defaulted to the lower levels, resequenced modules to match how learners would actually move through the form, and added scenario-based assessment types the AI had not suggested. The refined outline served as the blueprint for the Rise 360 build.
Source Material
Artifacts
Highlights
Scenarios, knowledge check
Tools Used
Rise 360 · Claude AI · TRACI prompting · ADDIE · 5 Moments of Need · Bloom's Taxonomy · Cavalier L3
Compliance Chatbot Simulation: By the Book
.png)
Click on the image to enter the simulation.
Challenge
Employees in regulated industries such as banking, healthcare, insurance, are often trained on compliance procedures in isolation, without ever practicing what it feels like to apply those procedures under real pressure. The hardest compliance moments aren't the ones where the right answer is unclear. They're the ones where you know the right answer but a trusted colleague is pushing you to move faster. This simulation puts the learner in exactly that situation: working through a routine end-of-period compliance checklist alongside Joe, a seasoned peer who is frustrated, blunt, and ready to be done for the day.
My Process
I used AI to help design every layer of the simulation, from the scenario concept and character personality to the behavior triggers, learner response guidance, and simulation ending conditions.
I defined Joe's emotional arc, the specific keywords that escalate or de-escalate his behavior, the range of acceptable learner responses, and the exact conditions that signal practice is complete. Each component was reviewed and refined to ensure the simulation measures correct process behavior — not scripted language or risk articulation.
The scenario was built in Devlin.ai and designed to transfer across any regulated field.
Artifacts
Highlights
AI-designed character simulation, behavior triggers, peer pressure scenario, regulated industry
Tools Used
Devlin.ai · Claude AI · TRACI prompting · 5 Moments of Need · Bloom's Taxonomy · Cavalier L3
