- Situation
- A 200-person digital mental health platform operating across the US, Canada, and UK. No one in the organization was using AI, there was no governance in place, and no shared way of working with it. In a regulated, care-driven environment, that meant both a missed opportunity and a real risk if people started using AI without guardrails.
- What Praxa did
- There was no budget, no mandate, and no one assigned to this. I started the AI enablement work myself because the organization needed it, and drove it across all 200 people. Rather than training people on a single tool, I taught the fundamentals that transfer across all of them: prompt engineering, how to spot hallucinations and bias, and the guardrails to put in place for safe use. That gave the team the confidence to actually pick up AI, with three tools introduced alongside the training to match the work: Perplexity, Cassidy, and Copilot. I wrote the AI policies for safe, responsible use across a regulated environment, and developed the organization's positioning for how it uses AI in service of its members. Alongside the enablement, I built an agentic compliance workflow that cut the time to complete customer compliance questionnaires from two weeks to three days, and the hands-on work on each from about 10 hours to 3.
- What changed
- The organization went from zero AI users to half the company using AI daily, with most of the rest using it regularly. Just as important, the culture shifted. A workforce that started with no exposure to AI became open and curious about it, including at the leadership level, where it became part of how senior leaders approach their own strategy work. The governance framework means all of this runs safely, with clear rules for a care environment where trust matters most.