[object Object]

Freshdesk supports 40+ languages. The hard parts are not the toggle. They are fallback rules, knowledge base translation parity, and where AI translation helps versus hurts.

Set the primary, then plan the fallbacks

The primary language defines the canonical content. Every other language is a translation that may lag. Configure the fallback chain explicitly:

Customer language: pt-BR
Solution lookup order:
  1. pt-BR
  2. pt-PT
  3. en (primary)

If any tier is missing, you serve English instead of a 404. Do not accept the default chain; spell it out.

Translate solutions in waves

Trying to translate everything at once kills momentum. Use the Knowledge Base analytics to find the top 25% most-viewed articles. Translate those first. Track translation completeness via the API.

GET /api/v2/solutions/articles/{id}/translations
// Compare modified_at per language; flag lag > 30 days

Freddy auto-translation: where it helps

Freddy AI offers in-conversation auto-translation. It is excellent for support agent ↔ customer when the agent is not native. It is risky for marketing copy and legal disclaimers. Restrict auto-translation to ticket replies, not knowledge base publishing.

Routing by language

Create a dispatcher rule that assigns tickets to language-specific groups. Detect customer language from the contact field, the email body (via Freddy), or the channel (a German chat widget defaults to de).

Don’t forget the canned responses

Canned responses are per-agent or per-group, not per-language. Create a parallel set per supported language and tag them with a lang_xx prefix for searchability.

What to do this week

Define the language fallback chain, run the top-articles translation list, restrict Freddy auto-translate to replies only, and ship language-prefixed canned responses for your top three languages.

[object Object]
Share