# An AI course finder with guardrails (draft)

CPR Finder · Laerdal Medical × AHA. Published 10 November 2026, updated 15 November 2026. Topics: AI, product design, trust.

*This piece is a draft. I’m still filling it in.*

Learners had to guess between BLS, ACLS, PALS and Heartsaver. I built an AI tool that asks a few questions and recommends the right one.

## The biggest pain: picking the wrong course

Employers say “get CPR certified.” Then learners face BLS, ACLS, PALS and Heartsaver, first time or renewal, online or in person. So they guess.

A wrong guess means a refund, a rebooking, a support ticket, or a learner who just gives up.

Career switchers often pick the course with the longest name, because it sounds safest. That’s usually PALS, the hardest one to pass. One learner spent their last $100 on a course, and then their employer didn’t accept that format. The training site had to refund them.

I brought the problem to the team with the evidence and got everyone on board to fix it. We set one goal: every learner books the right course the first time. Right course, more finished bookings.

## Which job needs which course

Nurses usually know they need BLS. The people who guess are everyone else. Electricians and park staff, for example, often need BLS plus first aid, because their job can mean using a bag-valve mask.

I mapped US jobs by sector to the CPR course each one needs. That map feeds the course finder.

![My research board: US jobs grouped by sector, each colored by the CPR course it needs, from BLS and Heartsaver to ACLS and PALS](https://neha-prasad.com/home/cpr-jobs.jpg)
*Which sector, which course: my research board.*

## How the course finder works

CPR Course Finder is RQI’s first AI tool for learners. It asks about your role and your job, then recommends the right AHA course.

My first version asked a lot: your role, what your job needs, whether you’d rather practice in a group or alone, your budget and your time. I wanted it to feel like ordering coffee your way. Then a product lead at RQI Partners told me the 3 questions she asks every learner on the phone. I rebuilt the finder around those.

1. **The learner answers a few simple questions**: What is your role? Do you need the card for work? Do you respond to emergencies at work? Do you need first aid too?
2. **AI reads the answers, inside guardrails**: Guardrails keep every answer to a real AHA course, in the right format.
3. **One clear answer**: For example: BLS plus first aid, self-guided at a CPR Verification Station, with about 45 to 60 minutes online.

## Getting it ready to launch

- I took it from problem to working tool myself: I researched it, designed it and built it, then had the developers test it.
- I’m working with Laerdal’s AI and Data team on guardrails, testing and data masking, so it meets our privacy rules before it goes live.
- It’s built to sit where learners start. The AHA team is interested in adding it to Atlas, the course booking flow on the AHA’s own website.
- My pitch: if the AHA embeds it, we embed it on CPR Finder, and distributors can embed it too. Then we all speak the same language, and learners can move between the AHA’s site and ours.

The course finder is getting ready to launch, and I’m leading it.

## What I learned

- Bring the evidence. Showing the team the pain in learners’ own words got everyone on board to fix it.
- The best questions come from the people on the phone with learners every day. Theirs beat my long list.

## Related

- [Swipe: keep what fits](https://neha-prasad.com/work/niyora/swipe.md)
- [Designing Moon: I tested the AI first](https://neha-prasad.com/work/niyora/designing-moon.md)
- [Privacy as the foundation](https://neha-prasad.com/work/niyora/privacy.md)

Part of [CPR Finder](https://neha-prasad.com/work/cpr-finder.md) by Neha Prasad. Live: https://neha-prasad.com/work/cpr-finder/course-finder
