In development · Early access by request
buildnwrite Assess
A Claude-powered assessment agent for Korean HR and L&D teams. It scores written work against your rubric, shows the evidence for every score, and sends uncertain cases to people.
What it does
buildnwrite Assess evaluates job-application essays, promotion essays and course assignments against the rubric your team already uses.
- Cites the evidence. Each score points to the sentence it is based on.
- Drafts individual feedback. Each submission gets its own feedback draft for a reviewer to edit.
- Escalates instead of guessing. Uncertain scores go to human reviewers.
- Keeps a reviewer queue. Reviewers see what needs a decision in one place.
It is built on the Claude API from Anthropic, with a focus on Korean-language long-form writing.
How it works
- Upload your rubric and submissions.
- Claude scores each item with evidence.
- Uncertain items go to reviewers.
- Export scores and feedback.
Background
The product grew out of grading pipelines the founder ran for corporate courses, and from coaching an enterprise HR team through a large-scale essay scoring run with Claude.
Status
buildnwrite Assess is in development. Early access is available by request. Email [email protected] to ask.
Made by buildnwrite, Seoul, South Korea.