How to Scan Korean Business Cards (명함) with AI

Korean business culture runs on the myeongham. It’s presented and received with both hands (or the right hand supported by the left), studied for a moment, and placed respectfully on the table — because it tells you the two things Korean business communication is calibrated by: the person’s organization and their rank.

Which is exactly why digitizing them badly is costly. Get the surname wrong, flatten the title, or lose the hangul, and your follow-up starts by undoing a bad first impression.

What makes Korean cards different from Japanese and Chinese ones?

The three CJK card families share the surname-first convention but fail scanners in different ways:

• Hangul is an alphabet, not ideograms. Unlike kanji, hangul can be read phonetically — but romanizing it consistently is the trap. The Revised Romanization system (official since 2000) coexists with McCune–Reischauer spellings and, more importantly, with personal preference: 이 is Lee on most cards but Yi or Rhee on others; 박 is Park or Pak.
• A handful of surnames dominate. Kim (김), Lee (이) and Park (박) together cover nearly half of Korea. Your archive will hold many Kims — the given name, company and title do the real identifying work, so they must be captured precisely.
• Titles are load-bearing. 대표이사 (CEO), 부장 (department head), 과장 (manager), 대리 (assistant manager) — Korean workplace address is built around rank. A scanner that flattens every title to “Manager” erases the information the card exists to convey.
• Bilingual layouts are standard. Korean on one face, English on the other. They’re one contact and should merge automatically.

The four-step workflow

1. Photograph in batches. Spread the evening’s myeongham on the table; a single photo can carry several cards. BizCardPro.AI detects each card, crops it, fixes the perspective and pairs front/back sides.

2. Let the AI keep both scripts. Each card becomes a structured contact: 김민준 stays in hangul with Min-jun Kim beside it, the surname correctly split, the phone normalized to +82 international format, and the address parsed into fields. If the card prints its own English spelling, the extraction uses it as ground truth instead of guessing.

3. Review what the confidence score flags. Every card carries a 0–100% score. Sort ascending and spend two minutes on the uncertain ones with the original photo shown alongside.

4. Follow up in Korean. Flag the contacts that matter and let the AI draft a short, properly courteous Korean follow-up — 존댓말 register, correct honorific forms of address — which you edit and send from your own mail app. (How follow-up works.)

How should Korean contacts be stored?

The same side-by-side rule that governs Japanese meishi and Chinese 名片: native script and romanization together, never one without the other. You’ll search “Kim” from a Latin keyboard on Tuesday and paste 김민준 님 into a greeting on Thursday. Storing both — with the title preserved verbatim — means both moments work.

Export follows the same principle: CSV, Google Contacts CSV and vCard all carry the split name fields and E.164 phones, so your Korean contacts survive every import intact.

Etiquette footnotes worth knowing

• Receive a myeongham with both hands, or the right hand with the left supporting your forearm.
• Take a moment to actually read it — rank determines how the conversation proceeds.
• Don’t write on it in front of the giver; add your context notes digitally after the meeting.
• Follow up within a couple of days, in Korean if you can — which is no longer a barrier when the first draft writes itself.