Building the AI That Builds Second Chances

What happens when the people building AI tools for workforce development are also people who would have been served by them? A note from inside my first month as Emerge Career's RFP Lead.

Liz Phillips is RFP Lead at Emerge Career.

Last week our team had something we call Claude & Tell — an internal show-and-tell where someone walks the rest of us through what they've been building with AI. It was my turn. I demoed the custom skills I've been writing in Cowork to automate the RFP work I now own. One of my colleagues said something I haven't been able to put down since: that's an Emerge journey story; you should write about it.

So I'm writing about it.

Two years ago, I applied for this job at Emerge Career. They said no.

How I got here

I spent five years in prison. While I was there, I went to college through Ashland University, took notes by hand, and wrote my essays on a tiny keyboard hooked into a tablet on weekends and at night.

I also had a tech job inside — Televerde, a company that operates call centers inside U.S. prisons. One of the only websites we could reach on the work terminals was Fortune.com. That's where I first read about ChatGPT. An officer walked past while I was doing my schoolwork one day and said, “Too bad you guys are in here doing school the hard way — now there's ChatGPT.” He meant it as a joke. I took it as a tip.

When I came home, I used the only tool I'd started to understand. I used ChatGPT to write my resume, to prep for interviews, to structure my thinking when I sat down to apply for jobs.

In March 2024, I applied for a state engagement role at Emerge Career. I used ChatGPT to draft my work trial. The role went to someone else — the candidate they hired had “more extensive experience in the government sector and RFP writing.”

I came back in July 2024 as a business development rep. In my first year on the team, I went from booking meetings to running the full RFP cycle — sourcing, qualifying, writing, pricing, submitting. I taught myself the things that, two years earlier, had been the reason I didn't get the role: indirect rates, cost narratives, MOAs. I helped win multi-year contracts in our key states.

This month, I was promoted to RFP Lead. The function I'd been told I wasn't ready for two years ago is now mine to run.

What I'm building now

I used to spend Sunday nights writing the Monday priorities brief by hand, working through our tracker line by line, deciding what to flag for the team huddle. I don't anymore.

Inside Cowork, I've built a set of custom skills that handle the parts of my job that used to live in a notebook and in my own head. Scoring new opportunities against our fit framework before we waste a week on a bad one. Pulling structured compliance checklists out of 80-page solicitations in about a minute. Drafting technical narratives from anchor language in our prior winning proposals across Massachusetts, New York, California, and Colorado. Deep research agents prospect employer partners and CBOs while I sleep.

What these skills replace is the part of the work that wasn't really the work — the formatting, the cross-referencing, the where did I save that section tax. What I get back is the time to do the actual job: building the right partnerships in the right states for the right opportunities, and making sure our proposals tell the truth about what we can deliver for justice-involved learners across CDL, HVAC, diesel tech, and construction trades.

A question worth sitting with

When we talk about who's using AI to solve workforce problems, we usually mean the people at the top of the org chart. What happens when the people building the tools are also people who would have been served by them?

I notice it in the small decisions. The eligibility gates I weighted heavily in our scoring framework came from places I personally know cost real time. The defaults I wrote for partner outreach assume our learners are workers, not statistics. None of that is unique to me, but it isn't accidental either. I'm one of a small number of people writing these kinds of skills who has worked from both sides of this work.

This July, I'll be in Washington, D.C. as a Jobs for the Future fellow, joining a workshop on workforce legislation across Massachusetts, California, and Arizona. The week ends with a day on Capitol Hill. I'm going because the people writing workforce policy are still mostly not the people the policy is for. That's a gap I want to help close.

The work is the same

Emerge Career trains justice-involved people for CDL, HVAC, diesel tech, and construction trades careers across Massachusetts, New York, California, and Colorado. I'm one of those people.

The first time I applied here, I was using AI to keep up with a world that had moved without me. I'm still using AI to keep up. The difference is that what I'm keeping up with now is the work — not the world.

The tools are different now. The work is the same.

— Liz Phillips, RFP Lead, Emerge Career

If you're an employer, workforce board, CBO, or funder thinking about what a partnership with Emerge Career could look like, we'd love to hear from you. Reach out through emergecareer.com.