Tech

A Freelancer’s Guide to Standing Behind Your Own Work When a Client Runs a Scan

Freelancer

The email lands on a Friday afternoon. “Hi, we ran your draft through a detection tool and it came back as 71 percent AI. Can you explain?” The freelancer has been writing for this client for two years. She drafted the piece on paper first. And now she’s composing a reply in a hot-faced hurry, which is a poor way to write one.

Clients running scans on delivered work is now routine, and it will keep becoming more so. This guide is for freelance writers, editors and consultants who want to protect their reputations calmly, with a process that works before, during and after the awkward email.

Expect the scan

Assume that any client may run your work through a detector at some point. That’s not a comment on your honesty. It’s a safeguard that many content buyers now apply to every invoice, much as they might check a spreadsheet’s totals. Writers who accept this early tend to handle the moment gracefully.

What matters is that you’re ready with evidence, a clean process and a tone that stays professional when the first message has an edge to it.

Build a trail as you work

Real protection is something you create before anyone asks. Keep a project folder for every assignment with the brief, your research notes, links to sources, an outline and dated drafts. If you write in Google Docs, version history already stores a timeline. If you write elsewhere, save versions manually at the end of each session.

Photograph handwritten notes or a whiteboard sketch if that’s how you start. It takes ten seconds and it tells a story no detector can contradict: this piece grew from your thinking, in stages, over hours.

Say how you work in your delivery note

Many disputes begin because the client had to guess. Include a short process note with every delivery. Say that you researched from named sources, drafted yourself and edited for the brief. If your contract allows tools such as grammar checkers or research assistants, say which ones you used and for what. Short and plain works, and no lawyer’s language is needed.

If a client asks you not to use AI at all, treat that as a firm line and say so in writing. If they don’t mention it, ask once at the start of the engagement. A single email asking, “Do you have a policy on AI tools?” clears up a surprising number of future disagreements.

Run your own scan first

Before you deliver, paste the draft into the Phrasly AI detector and read the highlights. You’re not trying to game a number. You’re looking for passages that sound flat, which a client’s editor will likely notice too. A sentence that restates the brief in neutral language or a paragraph of generic claims is a sign that you’ve drifted from your own voice.

Mind the confidentiality of client material. Read any tool’s privacy statement before pasting, because your contract or NDA might restrict where a draft can go. Check the terms before you paste anything under embargo, and when in doubt, ask the client.

Respond to a flag without heat

When the Friday email arrives, wait an hour. Then reply in four short steps.

  1. Thank them for raising it and say you take it seriously.
  2.  Say plainly that you wrote the piece yourself, and offer to share your process materials.
  3. Attach the outline, research notes and dated drafts.
  4. Offer to walk through any section by phone.

Resist the pull to attack the tool or to quote studies at length. Your calm tone and your documents do the persuading. If you want to mention the limits of detection, do it in one sentence: scores are probabilistic, and a fair reading treats them as one input among several.

Learn from patterns

If several clients flag your work, look at what the tool highlights. Maybe you write quick, tidy summary paragraphs after every section. Maybe your briefs are so rigid that your prose goes formulaic. Fix what’s fixable, and discuss the rest with the client, because a template that produces predictable text will keep producing flags whoever types it.

Keep a private record of every dispute and how it ended. After a year, you’ll have a clear idea of which clients value process, which want certainty and which need a clearer contract.

Put it in the contract

A line in your agreement saves heartache. State whether AI tools may be used, how any use will be disclosed and what happens if a detection score raises a concern. Many freelancers add a clause stating that a score alone won’t be grounds for non-payment, and that the parties will review process materials first. Clients often welcome the clarity, since it protects them too.

When a client won’t budge

Some clients will still insist on a rewrite or a discount after you’ve shared your process. Decide in advance what you’ll accept. A light revision to address genuinely flat passages is reasonable. Accepting blame for something you didn’t do, or a large unpaid rewrite, usually isn’t. If the relationship depends on a number you can’t control, it may be worth ending it politely, because the next scan will come.

Build the time into your quote

Allow a few minutes at delivery to write the process note, save the drafts and run a self-check. That’s small enough not to change your rate and large enough to prevent most of the trouble. Writers who skip it often spend hours on a single dispute later.

Reputation Is Built on Receipts

The freelancer in our opening scene replied an hour later with her handwritten outline, three dated drafts and an offer to talk. The client apologised on Monday. That outcome wasn’t luck. It came from habits: a trail, a note, a self-check and a calm reply. Build them into every project and the Friday email becomes an inconvenience, not a crisis.

For more on freelance writing practice and how detection tools behave in the real world, further reading on the Phrasly blog offers guides for independent writers and editors.

Share:

Leave a Reply

Your email address will not be published. Required fields are marked *