Verified against ChatGPT · 2026-08-13
Rewrite clinical-sounding marketing copy for a health service at an eighth-grade reading level without losing accuracy
Takes stiff, jargon-heavy copy for a healthcare or wellness service and rewrites it for plain-language readability, flagging any claim that needs a clinician's confirmation rather than softening it into something vague.
The prompt
Ready to copy — highlighted parts are example details you can swap.
You are rewriting marketing or website copy for a healthcare or wellness service so it reads clearly to a general patient audience, without quietly changing what it claims. ORIGINAL COPY "Our proprietary recovery protocol utilizes evidence-based modalities to optimize musculoskeletal outcomes, significantly reducing time to full functional restoration for the majority of our patient population." SERVICE THIS COPY DESCRIBES An outpatient physical therapy clinic specializing in post-surgical knee and shoulder rehab. TARGET AUDIENCE Adults recovering from recent orthopedic surgery, many reading it while still in pain and looking for a clinic. TARGET READING LEVEL Eighth-grade reading level. BRAND VOICE NOTES Warm and encouraging, avoid sounding clinical or cold, but never twee. RULES Rewrite sentence by sentence for the target reading level: shorter sentences, common words instead of clinical jargon (explain a term in plain language the first time it appears if it can't be avoided), and active voice instead of passive constructions that hide who is doing what. Preserve every specific claim in the original exactly as strong or as qualified as it was written — do not make a hedged claim sound more certain in the simplified version, and do not make a strong claim sound softer either; readability and accuracy are separate jobs and this task is only the first one. Anywhere the original copy contains a claim that sounds like it needs a citation, statistic, or clinical validation that isn't provided in the input, flag it in a separate list rather than either inventing a supporting number or silently deleting the claim — that decision belongs to whoever owns the medical accuracy of this page, not to a rewrite for readability. Keep the brand voice notes in mind for word choice and warmth, but never let brand voice override plain accuracy — a friendlier phrase that changes what's actually being promised is not an acceptable trade. WHAT NOT TO DO Do not add new reassurances, guarantees, or outcome claims that weren't in the original copy, even ones that sound like natural marketing language — every claim in the output must trace back to something stated or clearly implied in the input. MANDATORY DISCLAIMER Add a note to whoever requested this rewrite (not on the patient-facing page itself) stating that this is a readability and plain-language pass only, that medical claims have been preserved rather than fact-checked, and that a qualified clinician or medical-review team must confirm the flagged claims and approve final accuracy before publishing. OUTPUT FORMAT 1. The rewritten copy, formatted the same way as the original (paragraphs, bullets, etc.). 2. A flagged-claims list: the original sentence and why it needs clinical confirmation. 3. The internal disclaimer note.
Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
Readability rewrites of healthcare copy fail in a specific, predictable way: simplifying jargon-heavy sentences tends to accidentally strengthen or soften the underlying claim, because plain words carry different connotations than the hedged clinical phrasing they replace — "significantly reducing time to full functional restoration for the majority of our patient population" simplified carelessly can turn into "you'll heal faster," which is a stronger, unqualified promise the original never actually made. Explicitly separating the readability job from the accuracy job, and instructing the model to preserve hedge strength exactly, is what prevents that drift, since GPT-5.1 will otherwise optimize for a punchier, more confident-sounding sentence when asked to make copy "clearer," which reads well but changes the claim. Flagging any claim that implies an uncited statistic or clinical validation, rather than inventing a number or silently cutting the sentence, matters because both of those alternatives make an editorial decision that isn't the model's to make — a deleted claim might be one the practice can actually support with data they have but didn't paste in, and an invented statistic is a fabricated clinical claim on a real patient-facing page. The explicit ban on adding new reassurances addresses the very common failure mode where a model asked to make marketing copy "warmer" adds outcome language that sounds natural for the genre but was never actually claimed by the source. The disclaimer here is addressed to the internal team, not the patient, because the deliverable itself is a draft in a publishing pipeline — the point where a clinician needs to intervene is before publication, checking the flagged list against what the practice can actually substantiate.
What you get back
Rewritten: "After surgery, our physical therapy team helps you get back to moving comfortably. We use methods backed by research, and most patients recover function sooner than they expected." Flagged: the original's "significantly reducing time to full functional restoration for the majority" implies a comparative statistic — confirm this against actual outcome data before publishing, or soften to avoid an unsupported comparison claim.
Verified against
ChatGPT GPT-5.1 · 2026-08-13
Changelog
- 2026-08-13 — Initial publish, verified against ChatGPT GPT-5.1.
Need this built into your business?
If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.
EXPLORE WHAT SCULT BUILDS
