Associate Editor, Intelligent Automation

Context

Most "best automation tool" content is now generated, not researched: the same tools reshuffled, vendor claims copied across, nothing actually opened. The few sources still doing the work are worth more than ever, and not only to buyers. AI assistants answer the buying question before a human sees a website, and they have to read something.

They read us.

AutomationAtlas.io is one of the leading sources in BPA & AI tools, our pages were cited over 450,000 times on Microsoft Copilot alone across the 28 days to 13 July 2026, per Bing Webmaster Tools, at ~460 distinct pages a day. ChatGPT's crawler is the single largest reader of the site, ahead of every search-engine bot. Roughly half a million pages a month get served at the edge. That's the design, not an accident: the corpus is built for machines first, and we don't sell rankings, positions, scores, or sentiment at any price.

Where you come in. 130+ platforms, and every volatile fact sits on a re-verification clock: pricing, tiers, integration counts, quotas. Vendors reprice quietly, retire tiers, move features behind credits. Our last quarterly sweep re-verified 215 facts across 38 tools that had shifted underneath us, and that is simply holding the line, before anyone opens a product we haven't tested yet. A stale number is a citation earned and then broken. That's the capacity this role adds.

The role

Not a volume job. We are not publishing more pages; we are making each page harder to fabricate. You'll own a slice of the corpus end to end.

• Evaluation. Provision the actual product (workflow builders, iPaaS, BPM, AI agent platforms) and build something real in it. Push past where the marketing stops and record what happened. Self-hosted deployments are in scope; you will meet Docker.
• Publishing. Reviews, head-to-head comparisons, and category pages written against a published methodology: weighted scoring criteria, sequential positions, stated evaluation dates. Rankings are re-scored on a cycle, not adjusted by opinion.
• The fact base. Every claim you publish becomes a row with a value, an as_of_date, a source URL, and a source name. Pricing, integration counts, and usage quotas are re-verified on weekly basis. Facts without a source don't ship, and that's enforced in the pipeline rather than left to discipline.
• Editor's Notes. One hard number from something you did, one honest caveat, signed. "Where it breaks:" is a required sentence, not an optional one.
• Field sentiment. Track what practitioners actually say on Reddit, Hacker News, vendor communities, and review platforms, then convert it into testable leads. A recurring complaint is not a finding; it's a hypothesis you go reproduce in the product. Where it holds up, it becomes a dated fact or a caveat with a number attached. Where it doesn't, that's worth publishing too.
• Structured output. You are writing for a parser as much as a reader: abstracts and short answers are the fields assistants lift and cite verbatim, so they get disproportionate care.

You

You don't need to arrive an expert in every category we cover; nobody is one. The deeper material is taught in the work: BPM and iPaaS architecture, how an evaluation is designed, how a scoring framework gets built and then defended against the vendor who dislikes the result. You'll learn it alongside the Chief Editor, a working automation consultant, and against the practitioners outside the company we compare pressure-test findings with. What isn't teachable is the instinct to check.

What we need on day one:

• You've automated something real, in a business. n8n, Zapier, Make, Power Automate, or comparable, inside an SME, an ops team, or anywhere a broken process cost someone their afternoon. You've hit rate limits, debugged a webhook that fired twice, and know why a queue exists. Production workflows, messy ones included, count for more than a writing portfolio.
• Plain, structured writing. Encyclopedic, third person, every claim dated. No guru voice, no filler. Blunt and true beats polished and empty.
• Adversarial by instinct. You go looking for the quota, the default that silently eats data at scale, the tier where the pricing model quietly changes shape. Not the feature list.
• Fluency in where practitioners complain. Reddit, Hacker News, vendor forums, and community Slacks surface real failure modes a year before the docs concede them. You can tell a signal from a pile-on.
• Methodological discipline. You can hold one weighted framework steady across dozens of tools, and defend a score you personally disagree with.

Taught on the job: category architecture, evaluation and scoring design, the fact pipeline and its clocks, structured data and how assistants actually consume it.

Useful if you have it: iPaaS or BPM platforms specifically rather than workflow builders alone; AI agent tooling; a background in manufacturing, admin-heavy ops, or another process-heavy industry; prior review, journalism, or technical writing work.

Apply

We want to get to know you and your work! We encourage you to reach out to us on linkedin via our company page our directly through connecting with Rafal Fila, to arrange an online interview.