The one-paragraph answer
If you're hiring one role, one time, and you'll never look at this data again — ChatGPT, paste each résumé in, ask it to score, copy the answer somewhere. It works. ~$20/mo for Plus. Done.
If you're hiring more than once or you want the candidates to see anything other than silence — Lucida. ChatGPT is a smart intern who forgets you the moment you close the tab. Lucida is the workflow that remembers your last five hires, learns what predicted the good ones, and gives every applicant a feedback URL on your domain. Five things ChatGPT structurally can't do, listed below.
The five things ChatGPT can't do (no matter how good the model gets)
1. Remember your hiring decisions across conversations.
ChatGPT Memory exists but doesn't structurally retain "Vikas's last 6 hires scored 78, 82, 65 on dimension X; the ones at 65 didn't work out." Lucida does. After 6 hires /admin/insights tells you which rubric dimensions actually predicted your hires — and quietly retires the noisy ones from your next rubric. You become a measurably better hirer over time. ChatGPT keeps you exactly as good as you are today.
2. Give every applicant a feedback URL on your domain.
When a candidate applies through Lucida and you reject them, they get a private URL within 24 hours: their score, the dimensions, the cited evidence, what to do next. That URL says "powered by Lucida" but lives on your apply flow. Every shared report is a hiring-brand moment for you. ChatGPT can't see candidates — only the résumés you paste. The candidate side of the equation is invisible to it.
3. Be the apply form.
Lucida is a public surface. Candidates apply at your-company.lucida.team/c/your-company/careers/role-name. Their résumé is processed, the AI runs, and the candidate gets a structured experience — not a silent "we'll be in touch." ChatGPT is a chat window for you. It has no candidate-facing layer.
4. Hold up under an audit.
Every Lucida score records the prompt version, the rubric weights, the cited evidence, and the human who signed off on the shortlist. When the EU AI Act binds in August 2026, when NYC AEDT enforcement tightens, when India's DPDPA profiling rules clarify — that paper trail is what a regulator wants to see. ChatGPT's chat history is not a hiring audit trail; you'd have to reconstruct it manually for every candidate.
5. Cite the résumé.
Lucida quotes the candidate's actual words for every score: "Built our on-call rotation fairness framework — stopped the burnout spiral". ChatGPT can do this if you ask very carefully — but its default is to summarize, not to cite. The default behavior matters when you've scored 30 candidates and need to defend a shortlist to your co-founder.
Feature comparison
| Lucida | ChatGPT Plus (DIY) | |
|---|---|---|
| Price (entry) | ₹2,499 / month | ~$20 / month |
| Scores résumés against a rubric | Yes, structured, cited | Yes, if prompted carefully |
| Remembers your past hires | Calibration insights across roles | Conversation-scoped only |
| Public apply form | your-co.lucida.team/c/… | No candidate surface |
| Candidate feedback report | Automatic per applicant | You'd have to email manually |
| Audit trail (prompt version + evidence + human sign-off) | Per row, Save-as-PDF | Chat log, no structure |
| India-native (INR · GST · DPDPA) | Yes | USD · no GST · OpenAI ToS |
| Best fit | Founder hiring 3-6 roles/yr, wants to compound | One-off informal hire, never hiring again |
When ChatGPT is actually fine
We'll say it because it's true:
- You're hiring one informal role from your network — a friend's friend, a referral you've already vetted, where you just want a second opinion on the résumé before the call.
- You don't care what the candidate sees — usually because the candidate is someone you already know.
- You're testing whether AI scoring is even useful for you — paste a few résumés into ChatGPT first; if the scoring already feels valuable, then Lucida pays for itself when you hire your second role.
When ChatGPT stops being enough
- You're hiring more than one role this year.
- You've ever wished a previous candidate had received honest feedback.
- Your co-founder asks "why did you reject candidate X?" and you wish you had the cited evidence ready.
- You've ever wondered "am I getting better at hiring?" after a few months.
- You're starting to think about DPDPA / EU AI Act / AEDT before launching publicly.
Any one of those is the moment to move from ChatGPT to Lucida. They're all five reasons we built it.