From EA to Automation Lead: Real Paths Into a Field Most People Enter Sideways

Connor Sheffield had never worked a desk job. He didn't know what an API was. When he joined Zonos, a SaaS company, in 2021, his first assignment wasn't scheduling meetings or managing a calendar — it was keeping 15 live workflow automations running for the Chief Revenue Officer. He had zero automation experience and no formal training. He learned by comparing failed runs to successful ones, asking at every step what was breaking and why.

Three years later, he was Head of Marketing Operations and Automation. One system he built cut customer onboarding time in half. Another saved the executive team three to five hours a week, each.

Now meet Leon Petrou. Four years of industrial engineering at the University of Pretoria. His first workday happened to fall on his birthday. His first assignment was manually converting 10,000 addresses into geographic coordinates. One week in, he Googled "how to automate boring office work?" — and that search redirected everything. The same year he discovered RPA, he landed his first contract at triple his South African salary and handed in his resignation.

Two people. Neither had the credentials most readers assume are required. According to UiPath's 2022 survey of more than 1,200 working automation professionals, three in five transitioned from an entirely different role — and 53% described the transition as easy. The people already doing this work mostly started somewhere else. The question isn't whether someone with your background can make this shift. It's which version fits where you are right now.

But "automation specialist" isn't one job. Before the lane you pick determines what you study, it helps to understand what the work actually looks like from the inside.

What These People Actually Do at Their Desks

The core of automation work is not writing code. It's understanding what's worth automating, designing systems that survive real-world friction, and translating business pain into something that runs without supervision.

Connor's onboarding workflow at Zonos is a useful example. When a sales opportunity closed in Salesforce, data was cleaned and synced to the backend, a dashboard account was created, an onboarding project was spun up visible to both customer and rep, meetings were booked, carriers were connected, and the onboarding team was alerted. He didn't start by building any of that. He started by asking his colleagues, "What sucks about your job?" He documented the pain before he touched a single tool. The 50% reduction in onboarding time came from understanding the process, not from knowing the platform.

Antony Edwards, COO at Eggplant, put it plainly in an interview with The Enterprisers Project: "RPA is not just automating a spec. It's about understanding what the process is — and understanding the exceptions and stakeholder expectations." The analytical work precedes and outweighs the technical build.

Automation isn't about replacing humans. It's about elevating them to do what only humans can do.
— Connor Sheffield, Head of Marketing Operations and Automation, Zonos

That matters for career changers, because the skills that protect automation programs from failure — process documentation, exception logic, stakeholder communication — aren't engineering skills. An Ernst & Young study found that up to 50% of initial RPA projects fail, and the root cause is typically not the technology but how it's implemented. Bots break when data format assumptions change, when UI fields shift, when no one planned for the edge case.

If you've spent time in operations, project management, or any role where you had to explain why something kept breaking to someone who didn't want the technical details — you already carry the foundational literacy this work requires. A teacher who designed differentiated instruction for variable learners, a nurse who spotted intake bottlenecks, a recruiter who standardized a screening process: all of that contains the same analytical muscle the field actually values. The tool knowledge is learnable in weeks. The judgment about what's worth automating takes longer, and you may already have it.

Knowing what the work involves is step one. The harder question is which version of the work fits your background — because the salary ceiling, the credential requirement, and the time-to-readiness vary significantly depending on which lane you choose.

Four Lanes, One Field

There are four distinct paths inside "automation specialist," and which one fits depends less on how technical you are now than on what kind of work you're currently doing well.

Lane 1 is the developer path, building and maintaining automations in platforms like UiPath or Power Automate. UiPath puts the average U.S. starting salary at around $87,000, and Salary.com's April 2026 data puts the Junior RPA Developer average at $84,112 — with experience banding reaching $175,322 at eight-plus years. Best feeder roles: software development and QA testing. This is Leon's route. He completed a UiPath course, posted gigs on Upwork before he felt ready, and scaled into a consultancy.

Lane 2 is the business analyst path — identifying what's worth automating, writing requirements, governing the backlog. UiPath cites a starting salary around $80,000. No coding required. A Raymond James job posting for an automation intake analyst required five-plus years of business analysis experience; the UiPath BA credential was listed as "preferred," not mandatory. Best feeder roles: business analysis, process improvement, project management.

Lane 3 is the operator or internal inheritance path — Connor's route. You inherit or expand live automations inside your current employer. It's the lowest-friction entry point. The salary varies because you often start at whatever your current role pays, then grow from there. Best feeder roles: operations, service desk, executive assistance.

Lane 4 is the agentic-AI engineer path — combining large language model reasoning with workflow orchestration. UiPath cites a starting salary around $95,000. This is the fastest-moving and least-defined lane. No stable credential trail yet. Portfolio is the credential. Best feeder roles: data analysis, ML-adjacent work.

Choosing a lane isn't a minor detail. It determines which 10-week study plan is relevant and which hiring manager reads your resume as a fit rather than a stretch. If you come from HR, operations, or service management, Lane 3 doesn't require you to become a developer first. If you come from finance or accounting, Lane 1 is the fastest-growing application surface for your domain knowledge — Forrester estimates 36% of all RPA use cases originate in finance and accounting.

How Long It Actually Takes

The honest range depends on your path, not your pace.

Twelve to twenty-four months if you're on the internal inheritance track. You already work somewhere that uses automation tools. You volunteer to own one workflow this quarter, demonstrate ROI, and get re-titled. Connor's path. Fastest and least volatile because you're not convincing a new employer of your potential — you're showing an existing one.

Six to eighteen months if you go the credential-and-freelance route. You complete one vendor certification, publish your first automation project publicly, and take a contract before you feel fully ready. Leon's path. Higher income ceiling faster, but more uncertain in the early months.

This transition didn't come overnight. It took years to lay the groundwork for this career shift.
— T.J. Maher, SDET and Automation Developer

Eighteen to thirty-six months if you're studying while employed. T.J. Maher, a manual QA tester since 1996, described his own move into automation development this way: "This transition didn't come overnight. It took years to lay the groundwork for this career shift." He studied during evenings and weekends, used family support as his motivation, and credits a job audition — pitched as manual testing — as the bridge into writing Java and Selenium daily. He's been an SDET and Automation Developer since 2015.

One more reference point: Nathan Kahaialii, an elementary school teacher, began TripleTen's AI Automation program on January 1, 2025, while continuing to teach full-time. His LinkedIn headline reads "School Teacher turning into an AI Automation Specialist." That transparency — acknowledging the in-progress state rather than backdating a finished arc — is itself a signal of how this field actually works. Most transitions are visible while they're happening.

Worth noting: Microsoft retired its Power Automate RPA Developer Associate certification on June 30, 2026. Anyone who spent months studying for it after that date couldn't earn the credential. The lesson isn't that certifications are useless — it's that the landscape shifts fast enough that tracking vendor roadmaps is part of the job, even before you have the job.

If you have dependents, a mortgage, or a role you can't immediately leave, the eighteen-to-thirty-six month study-while-employed path isn't the slow option. It's the one designed for your situation.

What You Do This Week

Connor Sheffield's first instruction at Zonos wasn't "learn automation." It was "make sure these 15 Zaps don't fail." He didn't wait until he understood the system. He started with the broken workflows in front of him and learned by contrast, comparing what failed against what didn't. Three years later he runs marketing operations for a SaaS company. He still hasn't cited a certification.

The field isn't looking for people who already know how to automate. It's looking for people who can identify what's worth automating — and that judgment lives inside work you've probably been doing for years. The technical layer is learnable. The process instinct, the stakeholder fluency, the tolerance for ambiguity when a workflow breaks at 2 a.m.: those take longer to develop, and you may already have them.

This week, open a blank document and list three repetitive tasks in your current job that follow a consistent pattern — same inputs, same outputs, same decision logic each time. Don't research tools yet. Just document what the process is, what typically breaks it, and who else touches it. That document is the first artifact of an automation specialist's practice. Whether you take it to your current employer or to a portfolio project is a decision for next week.

The transition starts not with a course, but with noticing what's worth fixing — and you've probably been doing that longer than you realize.

Recommended Tools & Resources