Editorial policy

How we write and update the guides on this site, and the rules we follow when we edit your documents.

Who writes the guides

We publish the guides under our own name, without an individual byline, and a qualified editor reads each one line by line before it goes live. The topics come from problems we see in the documents we edit.

How guides are sourced

When a guide relies on a fact or a figure, we cite the original wherever we can (a style manual, a publisher’s guidelines, a piece of research, or the legislation itself) and link to it so that you can check it for yourself.

Keeping guides current

Every guide shows the date it was published and, if it has been revised, the date it was last reviewed. Our guides on AI-assisted writing rest on reasoning that should still hold as the tools change, and where one does depend on current facts, that date tells you how recent they are.

Our use of AI

We use software for mechanical work, including generative tools. Nothing goes live until a qualified editor has taken responsibility for it, and every position in a guide is one we hold. Client work follows the same principle: tools can help with mechanical tasks, but a person makes the editorial decisions.

Independence

We don’t run sponsored content, and nothing here has been paid for or placed by a third party. Nor do we earn commission from the tools and publishers we mention, so when a guide recommends something, it’s because we use it or would.

What we won’t publish

You won’t find case studies or client names on this site. When a guide draws on work we’ve done, it describes the pattern without identifying a document or an organization, and its worked examples are written from scratch, never adapted from a real document, even an anonymized one. We also stay away from guides on getting past AI detectors, since the detectors don’t work reliably.

How we edit your documents

Everything we return comes in tracked changes, so you can see each edit and reverse any you disagree with, and anything that affects your meaning or your argument goes in a margin comment for you to decide on. We don’t send clean copies with the changes accepted because working through the edits is an integral part of the job and keeps the writing yours. Before any work begins, you tell us what the document is and what it has to do, and we agree the scope and quote for it with you.

We check references as standard, for consistent formatting and to confirm that each one exists. Whether a source supports the claim it’s cited for is developmental work, so let us know if you’d like that checked and we’ll quote for it separately.

We turn down work meant to lower an AI detection score, and we don’t run documents through detection or plagiarism tools of any kind because we can’t see how those tools reach a result or verify what they report. Independent testing has found AI detectors to be neither accurate nor reliable (Weber-Wulff et al., 2023) and prone to false positives, especially for writers whose first language isn’t English (Gonçalves & Gonçalves, 2026), while plagiarism checkers can miss copied text and flag original writing (Foltýnek et al., 2020). Ghostwriting is for business clients only, since work submitted for a degree has to be the student’s own, and when we edit academic work, we stay within the rules set by your institution or funder.

No editor can promise publication, a grade, a won contract, or a funded application, and we don’t guarantee outcomes either. If we miss something material, though, let us know and we’ll put it right at no charge.

Your document stays confidential: only the people working on it see it, and we’ll sign an NDA if you need one. We keep files for a year in case you come back with follow-up work, and we can delete yours sooner on request. The FAQ covers most of these points in more detail.

References

  • Foltýnek, T., Dlabolová, D., Anohina-Naumeca, A., Razı, S., Kravjar, J., Kamzola, L., Guerrero-Dib, J., Çelik, Ö., & Weber-Wulff, D. (2020). Testing of support tools for plagiarism detection. International Journal of Educational Technology in Higher Education, 17, Article 46. https://doi.org/10.1186/s41239-020-00192-4
  • Gonçalves, B. F., & Gonçalves, V. (2026). Tools for detecting and humanizing AI in academic work: An exploratory study. In A. Rocha, C. Ferrás, & L. Loo (Eds.), Information technology & systems (Lecture Notes in Networks and Systems, Vol. 1965, pp. 588–597). Springer Nature. https://doi.org/10.1007/978-3-032-25652-2_48
  • Weber-Wulff, D., Anohina-Naumeca, A., Bjelobaba, S., Foltýnek, T., Guerrero-Dib, J., Popoola, O., Šigut, P., & Waddington, L. (2023). Testing of detection tools for AI-generated text. International Journal for Educational Integrity, 19, Article 26. https://doi.org/10.1007/s40979-023-00146-z

Seen something wrong?

Email hello@refineryediting.com with the page and the problem. We’ll check it and correct it if you’re right.

Get a quote