Sales Email Sentiment Annotation

Budget: €12 - €18/hr

I have a collection of outbound and reply-thread sales emails that need to be tagged for sentiment so I can train an internal automation model. The raw text is already cleaned and de-duplicated; what I now require is precise, message-level sentiment annotation (positive, neutral, negative).

You may work in Prodigy, doccano, Labelbox, or an equivalent text-annotation platform—feel free to suggest another if it exports clean CSV or JSON. Accuracy is more important than speed, so I’ll review an initial sample together with you before moving on to the full set.

Deliverables • A sentiment-labeled dataset of approximately 5,000 email messages in CSV or JSON format • A short read-me that explains any edge-case decisions or labeling guidelines you applied

Acceptance criteria • ≥95 % agreement with the validation sample I will check • File encodes properly in UTF-8 and opens without errors in Python/pandas

Once the pilot batch is approved we can extend the scope to further annotation tasks such as lead intent or key-point extraction, so clarity and reliability on this first round are essential. Looking forward to collaborating with you.