Automation Analyst

About the Role:
Most analyst roles end at the handoff: you produce the requirements doc, someone else builds it, and you find out six months later that what shipped doesn't match what the business actually needed.
This isn't that role. You do the analysis and you build the thing.
That combination used to be unrealistic to ask for. It isn't anymore. AI coding tools have collapsed the distance between understanding a process deeply and being able to automate it yourself, and we're hiring for exactly that shift. If you're an analyst who has spent the last year or two building things you couldn't have built alone before — because Claude, Cursor, or Codex closed the gap — this role is aimed squarely at you.
Here's where we are. We've already invested in automated process discovery tooling, so the raw picture of how work flows through this company is being captured for you. We also have a substantial portfolio of RPA automations in production that you'll help inherit. Some are excellent, some are brittle, some probably shouldn't exist.
Your job spans both ends. Discovery software shows what people do; it can't tell you why or which exception matters. That takes someone who will sit with users and ask the third and fourth question. Then you take what you learned and ship the automation yourself.
What You'll Do:
Interview and observe the humans doing the work. Run discovery sessions, workshops, and side-by-side walkthroughs. Dig past the first answer and surface the exceptions, workarounds, and edge cases that tooling alone won't reveal
Build the automations. Workflow orchestration, RPA, scripts, integrations, AI-assisted steps — you're expected to deliver working automation, not just specify it. You'll pull in engineering for the genuinely hard problems, not the routine ones
Turn process discovery output into a prioritized, quantified backlog: what it costs today, what it costs to fix, what we get back
Help own our existing RPA portfolio: understand what each automation does and for whom, fix what's brittle, document what's undocumented, and retire what no longer earns its keep
Write the requirements: user stories, acceptance criteria, exception paths, and the failure modes everyone forgets until go-live
Own reliability for what you ship: monitoring, error handling, and runbooks
Manage the change: bring affected teams along, train them, and make sure the new process actually sticks
Measure what shipped against what you promised, and report it honestly
What We're Looking For:
You build, not just document. Point to automations, tools, scripts, or workflows that exist and are used. We care more about what you've shipped than what you've specified
Serious hands-on AI tooling experience. You work regularly in tools like Claude, Claude Code, Cursor, or Codex — using them to write and debug real working code, not to draft emails. You know how to decompose a problem for a model, when to trust the output, and where it will confidently hand you garbage
Real analyst chops. You can run a discovery session, elicit requirements from someone who's never written one, and turn a vague complaint into a defined problem with a number attached. You ask the follow-up question
Technical fluency. REST APIs, JSON, webhooks, SQL, and scripting in Python or JavaScript. You should be able to read an API doc, authenticate against it, and get data moving without help
You automate your own work. You've built the script, the macro, the workflow that saved you three hours a week — and then you built four more
Genuine curiosity about how work gets done. The best insight in this job usually comes from watching someone's screen for twenty minutes, not from reading a process doc
Willingness to maintain, not just build. A meaningful share of year one is understanding and improving someone else's work. If that sounds like a chore rather than a puzzle, this isn't the right fit
Systems thinking about people. You can talk to a controller about their month-end close, understand the edge cases they're worried about, and get buy-in before you change how their job works
You know the difference between a process that should be automated and one that should be eliminated or renegotiated first
Bonus Points:
Business process modeling — BPMN or equivalent, at a level of detail people actually read
Hands-on with workflow automation or iPaaS platforms
RPA experience, especially inheriting or auditing an existing bot portfolio. n8n experience.
Built something real with LLM APIs: structured extraction, document processing, tool-using agents, evaluation harnesses
Familiarity with task or process mining tools and translating their output into action
Formal continuous improvement background: Lean, Six Sigma, value stream mapping
Browser and UI automation: Playwright, Puppeteer, Selenium
Data work beyond SQL: dbt, warehouse modeling, BI tooling
Why This Role?
You own the whole loop: find the problem, size it, design the fix, build it, prove it worked
We've already invested in discovery tooling, so you spend your time on solutions instead of hunting for problems
Real scope from day one: a live portfolio to help own plus a mandate to build new
Budget and mandate for the tools you need, including AI tooling. We want you using it aggressively
Direct line to leadership and to the people whose work you're improving
Benfits & Perks:
Health / Dental / Vision Insurance Plans
401K Plan with Employer Match
Life Insurance Plan (100% employer paid)
PTO Plan (3 weeks PTO for new employees, grows with tenure)
6 paid holidays
Paid Parental Leave
Free professional coaching through Boon Health
Baby-feeding support and on-site Lounge through Pumpspotting
Comprehensive training program
Commitment to defend employees who have signed predatory non-competes (Learn more here)
Ongoing, in-house professional development
In-office gym complete with strength and cardio equipment as well as Peloton bikes
On-site snack machine

Employment type
Full-Time

Work arrangement
No