An always-on front door for a luxury salon that meets each customer in their own language, answers the questions stylists were fielding all day, and guides them confidently toward a booking.
BeforeEnglish onlyThe same price, preparation, duration and suitability questions across phone, social and web, with non-English and after-hours enquiries slipping away
After12+ languagesAnswered automatically in the customer's own language, day and night, without pulling a stylist off the floor
01 · Overview
Stylists back on the floor.
A luxury salon specialising in braids, protective styles, extensions, silk press and curly hair care was burning a meaningful share of every stylist's day answering the same handful of questions. A substantial part of the clientele spoke languages other than English, and every English-only reply slowed the conversation or dropped the enquiry entirely.
NFI built Clara, a conversational assistant that answers service, pricing, preparation and duration questions across more than twelve languages, detecting and responding in the customer's own language automatically.
02 · The problem
The same questions, in a dozen languages, at all hours.
Price, preparation, duration and suitability for different hair types, asked across phone, social media and the website, with after-hours messages losing their momentum overnight.
01
Stylist time on the phone
Answering routine enquiries pulled stylists away from the chair for a meaningful share of every day.
02
Lost in translation
English-only replies slowed non-English conversations down or lost the enquiry entirely.
03
Momentum lost overnight
After-hours messages waited until morning, by which time the customer had often moved on.
03 · The solution
Reasoned answers, in the customer's own language.
01
Automatic language detection
Clara detects the customer's language and replies in it, across more than twelve languages, with the same accuracy on service details in each.
LayerLanguage
02
Private reasoning step
Tricky compatibility and preparation-versus-inclusion questions are worked through privately before Clara replies, so what is included, what to do in advance and which services suit which hair types are answered reliably rather than guessed.
LayerReasoning
03
Cost and uptime in check
Timezone-aware generation keeps availability and greetings aligned with salon hours, and aggressive prompt caching combined with a multi-model fallback chain keeps both cost and uptime under control, with full observability in production.
LayerResilience
04 · The results
Always on, in every language the clientele speaks.
12+Languages supported for enquiries and responses, with automatic detection
~90%Prompt cache hit rate, delivering measurable per-request cost savings
10Model fallback chain supporting resilient uptime across providers
FreedStylists back with clients instead of on the phone, with full observability in production