Home / Case studies
Production, not prototypes Selected work

AI systems that changed real operations

A selection of platforms and AI agents running in production. We show the operating problem, what we built, and the outcome the client could actually measure.

AI video SaaS

Feberra — video production from a script in two minutes

Professional video used to mean a camera crew, studio time, editing, and a budget of $500 or more per asset. We built a self-service SaaS product that turns text into presenter-led HD video for global audiences.

What we delivered
  • End-to-end text-to-video workflow with automated rendering and subtitles
  • 50+ languages and 100+ AI avatars for localized campaigns without reshoots
  • Stripe subscriptions and scalable rendering infrastructure for a repeatable SaaS business model
2 min
Average video creation timepreviously: 10 hours of editing
10K+
Usersorganic growth in the first months
50+
LanguagesCIS, European and Asian markets
CRM-native AI agent

TermoLand — an AI manager handling 17,000 messages a month

A growing thermal spa network was losing enquiries after hours and during call-center peaks. We embedded “Anna” directly into the client’s CRM, so AI and human operators work in one conversation history and hand-offs keep the full context.

What we delivered
  • 24/7 consultation and lead follow-up inside the existing CRM workflow
  • Multi-model architecture balancing response quality, speed, and token cost
  • Human escalation and quality-control loop for uncertain or sensitive requests
17K
Messages per monthhandled by AI
99%
Answer accuracyerrors below 1%
2 min
Response timepreviously: hours at peak
80%
Deferred dealsfollowed up by AI
AI manager Anna follows up with a customerHuman operator and AI response comparison
Alexander Polkovnikov, CPO at TermoLand: “We used to drown in call-center turnover and missed enquiries — and lose customers. Now Anna, our AI bot, handles it quickly, accurately, and without fatigue.”
WhatsApp sales agent

Optical retailer — appointment conversion increased from 10% to 32%

Leads were waiting up to 40 minutes and leaving for competitors. The AI agent now answers on WhatsApp, qualifies each enquiry, checks availability, books appointments, and hands complex cases to a manager with full context.

32%
Lead-to-appointment conversion17 bookings from 53 leads; previously 10%
19
Appointments in one daya new store record
70%
Bookings made by AI12 of 17 without a manager
Client feedback: “Conversion is higher. Of 17 appointments, five were mine; the bot booked the rest.”

Have an operational bottleneck AI could remove?

Tell us what is slowing your team down. We will map the workflow and propose a practical first implementation.

Discuss your project