Claude is a better starting point than ChatGPT for LPN resumes, but it has its own failure mode: it’s overly cautious about medication and clinical claims. Where ChatGPT overstates your scope, Claude understates it. It hedges valid scope-of-practice details, adds qualifiers like “as directed” or “under supervision” when you actually performed tasks independently per RN care plan, and sometimes removes medication routes or clinical details because it’s unsure whether LPNs can do them in your state.
The result is an LPN resume that sounds less capable than you actually are. Here’s how to use Claude effectively while fixing its cautious output.
The core problem: Claude over-hedges LPN scope
Claude is trained to be cautious about clinical claims, which is generally a good instinct. But on an LPN resume, this caution becomes a liability. Specific failure patterns:
- Adding unnecessary qualifiers. “Administered medications as directed by the supervising RN” when you should say “Administered PO, IM, and SQ medications to 25 residents per med pass.” The “as directed” makes you sound like a CNA with a medication aide credential, not an LPN with independent med-pass authority.
- Removing medication routes. Claude sometimes replaces “PO, IM, SQ” with just “medications” because it’s uncertain about route-specific scope. Those routes are exactly what the DON needs to see.
- Softening wound care claims. “Assisted with wound care procedures” instead of “Performed wound assessments and dressing changes (wet-to-dry, wound vac) per RN care plan.”
- Dropping delegation language. Claude sometimes removes bullets about supervising CNAs, possibly because it’s unclear about LPN delegation authority in your state.
The prompt structure that works
Like ChatGPT, Claude needs explicit state scope context. But Claude also needs a confidence instruction — tell it not to hedge:
I'm an LPN licensed in [STATE]. In my state, LPNs are authorized to:
- Administer medications via [PO, IM, SQ, IV if applicable]
- [Perform focused assessments / collect data for RN assessment]
- [Start peripheral IVs / monitor IV infusions — if applicable]
- Perform wound care per RN care plan
- Supervise and delegate to CNAs
Do NOT add hedging language like "as directed," "under supervision," or "assisted with" unless I specifically used those phrases. I performed these tasks within my LPN scope — write the bullet with confidence.
My facility: [TYPE, e.g., 150-bed SNF]. My EHR: [PointClickCare]. Patients per med pass: [25].
Rewrite this bullet to sound like it was written by a confident, working LPN. Keep medication routes, patient count, EHR name, and all clinical details.
Original: [paste bullet]The anti-hedging instruction is the key difference from the ChatGPT prompt. Claude follows instructions well — if you tell it not to hedge, it usually won’t. But you need to say it explicitly.
A real before-and-after
The manual edit pass: what to check
After Claude produces output, check for:
- Hedging language. Search for “assisted,” “supported,” “as directed,” “approximately,” “under supervision.” If you did the task within your LPN scope, remove the hedge.
- Missing medication routes. Claude sometimes drops PO, IM, SQ in favor of generic “medications.” Add routes back.
- EHR specificity. Check that it kept PointClickCare, Epic, eClinicalWorks by name.
- Delegation bullets. If Claude removed your CNA supervision language, add it back. Delegation is a core LPN differentiator.
- Exact patient counts. Claude sometimes rounds (“approximately 25”) when you should be precise (“25”).
Where Claude beats ChatGPT for LPN resumes
Claude is better at: following formatting instructions (it won’t restructure your resume without being asked), maintaining your original voice, refusing to fabricate metrics you didn’t provide, and handling the LPN/LVN naming distinction correctly when instructed. The edit pass is usually faster because you’re adding confidence back to accurate claims, not removing fabricated buzzwords.
Frequently asked questions
Is it obvious to recruiters when an LPN resume was written by Claude?
Less obvious than ChatGPT, but still detectable. Claude’s tell on LPN resumes is over-hedging — phrases like “assisted with medication administration” or “supported wound care procedures under supervision” when you actually performed those tasks within your scope.
Should I paste my whole LPN resume into Claude?
One section at a time produces better results. Working section by section lets you verify that each rewrite preserves your medication routes, patient counts, and scope-of-practice claims.
Should I include my LPN license number in the Claude prompt?
No. Use a placeholder and fill in the real number in your final document.
Is Claude better than ChatGPT for LPN resumes?
For specific tasks, yes. Claude is better at formatting, voice preservation, and not fabricating metrics. But it over-hedges clinical claims. Neither tool produces a finished LPN resume without a manual edit pass.
How long should this process take per LPN job application?
Expect 20–25 minutes per application. Similar to ChatGPT, but the edit pass focuses on removing hedging language rather than removing buzzwords.