# AI Took My Job (And I Helped!) In 2021 I got hired as employee number thirteen at a tiny startup called Waltham Clinic. My official title was “medical assistant,” but that was more of a polite suggestion than a job description. In reality, I was a medical assistant / data entry guy / insurance verification specialist / behavioral health assistant / random-firefighter-of-everything. If something needed to get done and no one knew whose job it was, it somehow slid onto my plate. The company was scrappy. We were a handful of people trying to build this new model of care, figuring out processes as we went, duct taping together healthcare, insurance, and software. I was sitting at this weird intersection of clinical work and paperwork, talking to patients and providers on one side and dealing with insurance companies, forms, and endless spreadsheets on the other. --- ### From Clinic to “Real Startup” Then we rebranded. Waltham Clinic became Lightyear Health, and the whole vibe shifted. Suddenly we weren’t just a little clinic; we were a “healthcare startup.” Funding came in. New people started appearing in Zoom squares every week. We went from thirteen people to over a hundred. There was a shiny new CEO. There were departments. There were org charts. There were acronyms for everything. And there I was, still technically the medical assistant, but doing work that had almost nothing to do with taking vitals or rooming patients. I was doing billing and coding, helping audit claims, filing prior authorizations, handling requests for payment, and trying to make sense of all the ways money flows (and leaks) in American healthcare. My little corner of the company — the insurance team — was five people: me and four amazing ladies who knew the insurance grind inside and out, but didn’t consider themselves “technical” at all. That’s where things got interesting. Because while my title said medical assistant, my brain was acting like a systems engineer. I couldn’t stop seeing patterns in the chaos. We were doing the same steps over and over, manually, across twenty tabs and three different systems. Every prior auth, every verification, every claim felt like a workflow begging to be automated. And sitting right next to us (virtually) were the software engineers. --- ### The Grey Dashboard So I started talking to them. Not in fancy tech language. I just described what our day looked like, step by step. “First we log into this portal. Then we copy this ID. Then we check eligibility here. Then we paste it there. Then we screenshot this page for proof. Then we manually type the result into our internal notes. If we forget one of those steps, the whole thing breaks three weeks later when a claim denies.” The engineers listened. They asked questions. And then they started building. What came out of those conversations was our internal grey dashboard — the backend tool for the insurance team. It was not pretty. No one was going to dribbble for design inspiration off this thing. But for us, it was magic. Suddenly, instead of ten different clicks, we had one place to work from. Data flowed in from APIs. We could see what we needed without constantly logging in and out of third-party portals. We iterated. We filed tickets. We said, “Can you add a field for this?” or “Can it show us this status without having to click three more times?” That was all before AI showed up. --- ### Discovering AI While Everyone Ignored It In 2022, when AI started going mainstream, I got curious fast. I didn’t jump straight into coding — I used it to publish books. I would feed ideas into these early models, clean up what came out, and push it to Amazon. It felt like cheating, but also like unlocking a superpower I didn’t know I had. Back at work, though, nobody wanted to hear about AI. I’d mention it on calls: “Hey, there’s this thing that can help us draft emails, or summarize notes, or write appeal letters faster.” The reaction was… polite silence at best. People were busy. Legal was cautious. Healthcare is slow to move, and everyone was already overwhelmed by existing tools. So I kept my AI experiments mostly as an after-hours thing. Meanwhile, inside the company, I kept doing what I had always done: helping the engineers understand what the insurance team needed. I wasn’t writing code, but I was translating pain points into requirements: “This step right here? This is where our claims die. If you can automate this, you’re saving us hours every week and thousands of dollars every month.” One of the biggest wins we pushed for was integrating an API to automate insurance verification. --- ### The Automation I Helped Build If you’ve never had the pleasure of verifying insurance manually, imagine calling the same unhelpful phone number a hundred times a week while flipping through an endless binder written in legalese. Or logging into a dozen different payer portals that all look like they were designed in 1998. Every patient, every visit, every time, you’ve got to confirm eligibility, benefits, sometimes deductibles and copays, and log it all properly so the claim doesn’t bounce. We worked with the engineers to hook into a service that could do a lot of that checking automatically through APIs. Suddenly what used to take several minutes per patient could be pulled in seconds. We built queues, rules, and views so the team could focus on exceptions instead of doing the same repetitive lookups over and over. It was a huge win. It was also the beginning of the end for my job. Because if you zoom out, what was I really doing for the company? I was manually executing workflows that I was simultaneously helping to automate. Every time we improved that backend dashboard, we were quietly making my own role less necessary. I was helping build the machine that would eventually do my work better, faster, and cheaper than I could. I didn’t see it clearly at the time. I was just excited to fix broken processes and make life easier for the team. I liked being useful. I loved that the CEO and the original core team trusted me. I loved being the bridge between the front lines and the builders. --- ### When the Music Stopped Then things shifted. Markets got weird. Priorities changed. We pivoted offerings. We went from hiring like crazy to tightening budgets. The team shrank from over a hundred people down to around fifty. Leadership changed again. Projects got cut. Whole departments disappeared. Suddenly every role had a giant invisible question floating over it: “Is this critical to the new version of the company?” One day in September 2023, I found out the answer for my role. They let me go. Not because I was doing a bad job. Not because people didn’t like working with me. But because, in their eyes, there just “wasn’t anything else for me to do.” The workflows I had helped streamline and automate were working. The dashboard was built. The API was plugged in. And without a bachelor’s degree in computer science, I didn’t fit neatly into any of the higher-paying technical roles that survived the cuts. On paper, I was still just a medical assistant. In reality, I had been doing systems work, operations design, and informal product for years. But job titles are what get scanned by HR software. Job titles are what recruiters filter by on LinkedIn. And “medical assistant” doesn’t tell the story of someone who helped design automation that replaced chunks of his own job. I walked away with a strange mix of emotions: frustration, gratitude, fear, and this weird, simmering sense of possibility. I had just gotten a front-row education in how startups work — how money flows, how decisions get made, how quickly priorities can flip, and how automation is always quietly on the roadmap, even if people don’t say it out loud. --- ### From “Medical Assistant” to AI-Obsessed CS Student I also had something else: a growing obsession with AI. After I got fired, I didn’t rush into another job. I went back to school to finish my computer science degree. I went deep into AI. Not as a casual user, but as someone who tests the limits every single day. I started learning to code properly. I stopped just “talking to the engineers” and started becoming one. Here’s the wild part: the grey backend dashboard we had spent months designing at Lightyear? Today, I can build something better — more polished, with a real database, cleaner UI, and smarter automation — in a weekend with AI as my co-pilot. The thing that used to take a whole engineering team weeks of planning, stand-ups, tickets, and demos can now be prototyped by one determined person with the right prompts, some curiosity, and a willingness to iterate. That blows my mind. I’ve lived both sides of it: I watched humans build this stuff without AI… and now I can rebuild it faster with AI. So yeah, in one sense, AI “took my job.” But here’s the nuance: I helped it. I helped document the workflows. I explained the edge cases. I pointed out the steps that were fragile and the ones that didn’t need a human brain. I pushed for the API that eventually made my manual insurance verification work obsolete. Whether it was automated with classic scripts, fancy APIs, or next-gen AI, the direction was always the same: fewer humans doing repetitive tasks. --- ### Hard Truths and Better Questions That’s the uncomfortable truth we don’t like to admit: if your job is basically a series of predictable steps that can be written down, it’s on the chopping block. Sooner or later, someone — or something — will come along and do it faster. So what do you do if you’re in that position? Here’s what I’m doing, and what I’ve learned the hard way: 1. **Be the bridge, not the bottleneck.** The most valuable people in a company are often the ones who can translate between the business, the front line, and the technical side. If you can explain complex workflows in plain language and help turn them into systems, you’ll always be near the center of important conversations. 2. **Learn how the money moves.** Once you’ve seen how revenue turns into headcount, software budgets, and marketing spend — and how quickly it can all be cut — you stop taking any single job personally. You start seeing your role as part of a larger system that needs to make financial sense. 3. **Don’t cling to your job description.** Official titles will always lag behind what you’re actually capable of. Stretch beyond them. Take on projects that feel “above” your title. One day, that’s the story you’ll tell when you pivot careers. 4. **Treat AI like a power tool, not a threat.** You can sit there hoping it doesn’t change your job, or you can grab it and start using it to do ten times more than you could before. The people who lean in now are going to be the ones designing the next wave of workflows, not just surviving them. 5. **Get something formal on paper.** Degrees, certificates, portfolios — they matter. I wish I’d had my CS degree finished when the layoffs happened. That’s part of why I’m back in school now. Experience is gold, but the gatekeepers still scan for credentials. --- ### So… Did AI Take My Job? I joke that AI took my job, but really, my job was always destined to evolve into something more automated. I just got to watch it happen up close. And I’m glad I did, because it forced me to level up. Today, I’m not the guy who gets handed a grey dashboard and told, “Use this.” I’m the guy who can sit down with AI, sketch out the workflows, and actually build the dashboard. I’m not just explaining problems to engineers anymore — I’m becoming one of them. If you’re in a role right now that feels repetitive, fragile, or easily explained step-by-step, don’t wait for someone to automate you out of it. Start learning the tools. Start mapping the workflows. Start moving yourself from “the person who clicks the buttons” to “the person who designs the system.” AI might take your job as it exists today. But it can also hand you the tools to build the next one.