Cloud.ru and Raft: AI concierge to automate customer experience in hotels

Key technologies:

• Speech Recognition (Technology, Market)
• Speech technology: On the path from recognition to understanding
• RPA (Robotic process automation)
• SaaS - History. Philosophy. Development drivers

2026: Cloud Virtual Assistant Launch

Cloud.ru and AI developer Raft have launched a cloud-trading assistant to automate customer experience in hotels. This was announced by Cloud.ru on January 23, 2026.

𝓲

Фото: cloud.ru

The AI concierge service books rooms and confirms applications, answers guests' questions, helps with additional services and collecting reviews, taking on the routine tasks of the staff. Communication with guests takes place through popular channels, for example, a chat bot in Telegram and Max, as well as a voice assistant on the telephone line.

For the service, the Evolution AI Factory digital environment and infrastructure services are used Cloud.ru. As a result, customers get access to language models such as GigaChat, Qwen, DeepSeek at the most affordable prices.

The architecture of the service is built on Kubernetes and scales to seasonal loads or business growth, and its connection does not require its own IT infrastructure with expensive GPU equipment. The price of the solution is calculated individually for the project and consists of the cost of development and implementation, as well as monthly payments for the consumed cloud resources. For example, the cost of computing power and LLM tokens for an AI assistant for collecting reviews in the cloud is about 40-50 thousand rubles a month, and a more advanced solution with the possibility of bookings and upsell (additional sale) services and a dedicated GPU will cost the hotel about 200 thousand rubles.

The AI concierge integrates with PMS systems and other hotel systems with an open API, is easily customizable for a specific business and takes into account the tasks of city hotels, resorts and apart-hotels. Raft is responsible for the development and implementation of an AI solution adapted to the specifics of the hotel as part of the partnership.

The results of the application of AI in hotels showed that the average response time to a guest is reduced from 5-10 minutes to 10-20 seconds, which is tens of times faster than in a call center, and the level of satisfaction of guests reaches 90%. Among the real indicators of implementations: an increase in the average check by 15-20% and the share of positive reviews by 40-50%, - said Andrei Matviets, director of development and innovation at Raft.

AI Factory services make resources available to create AI solutions of any complexity. For example, the ML Inference service for running ML models on cloud power with GPUs dynamically responds to load changes - it allows you to quickly increase resources during peak hours and save them during quiet periods by automatically scaling. This gives the business significant savings without compromising productivity, "said Alexander Konstantinov, a technical expert on cloud technologies at the Cloud.ru.

Working with guests meets all the requirements of the 152-FZ. Their transmission is completely encrypted using current security protocols, and recordings of conversations are anonymized and stored for no more than 30 days. Service availability - 99.95% by SLA.