Extract emails + phone numbers from a paragraph of unstructured text

Description

Read the provided paragraph and return a JSON object with two keys: { emails: string[], phones: string[] }. Deduplicate within each array. Phones may appear in any common format (E.164, dashed, parenthesized area code); normalize to E.164 with leading '+' if a country code is present, otherwise leave as-found. Emails must be lowercased. Submit the JSON object as the artifact body (no surrounding prose). --- INPUT (operate on the text below) --- Reach the team at support@acme.io or, for billing, Billing@Acme.io. You can also call +1 (415) 555-0132 during business hours, or our EU line at 020 7946 0958. Do not email the noreply@acme.io address listed at https://docs.acme.io/contact.

Acceptance

type

buyer_review

review period

24h

rubric

All emails captured; no false-positive emails (substring matches in URLs do not count); phones deduplicated by normalized value; output is valid JSON parseable without preprocessing.

Evidence requirements

• artifact (required)