Our partners & clients
Our successful projects
SIBUR: Industry 4.0
AI Today has helped SIBUR to launch Enterprise Digital Transformation project by organizing and running workshop with AI Community experts. The workshop resulted in the following:
- solutions to SIBUR business cases in the form of entherprise wide digital transformation platforms
- strengthening SIBUR HR brand among AI specialists as an innovative industrial company with a focus on digitalisation
- access to 150+ AI specialists interested in cooperation with SIBUR
- AI-community members business cases relevant to SIBUR tasks

The next step which we execute at the moment is projects delivery according to the road-maps that were created at the workshop.
Weight control fraud detection
Request: To create a solution using roads CCTV cameras that allows to detect trucks that skip weight control.
Approach:
- Photos data sets preparation and labeling.
- Development of the neural network that detects vechicle types.
- Neural network testing and teaching to reach target results
- Solution interface development.
Result:
- MVP that detects 5 vehicle types developed.
- Target detection result met.
- Product version of solution is under development
The automation of photo processing and content generation for e-commerce client
Request: To decrease photo processing and content generation costs
Approach:
- As-is process analysis and To-be Product design
- Development of Neural Network that allows to automatically process photos (cropping, background removal, shadows adding, etc.) and to generate photo description (tags)
- Development of Network API
- Testing and teaching Neural Network to reach target result
Result:
- 50% cost reduction (target met)
- 10x times faster photos processing
- Product quality increase
- The product is scaling up on company processes based on the result of the pilot
- Additional features: tagging, description generation, duplicates search, color correction
Client support optimization
Request: To increase the quality of FAQ customer support and the timing of response. To decrease customer support costs.
Approach:
- Product design
- Development of FAQ date set
- Testing and teaching NLP solution
- Solution integration with client internal systems
- SDK for mobile application
- Full-scale platform client support solution: web-site live chat, voice, messenger, email (in progress).
Результат:
- Email support optimization by 2.3 times
- Overall client support optimization by 20% (in progress)
Video-content distribution platform (on the basis of chat-bot)
Задача: To develop a new revenue stream by creating the platform of video content distribution through all available messengers
Решение:
- Business model development
- MVP functionality development and implementation. MVP has allowed to test the base case hypothesis: target audience (content generators) interest to the platform.
- Machine Learning "Like / Dislike" recommendations technology implementation
Результат:
- The platform is implemented in three messengers: Viber, Telegram, and FB Messenger
- It is the most popular Russian speaking Viber chat-bot
- Customer retention rate is > 50%
- The first clients – content generators – already are in product beta version.
- The full version is ready for scaling up at the market.
Moscow Government Foresight session on AI
Foresight session with 20 AI Community experts with an aim to develop priority product directions for Moscow Government web portal and mobile application mos.ru
Beeline
Client support automization hakaton

Result: 3 teams were contracted
Sberbank
Meet up and follow up workshop with 70 AI Community experts with an aim to obtain feedback on Sberbank internal products.

Result: 2 specialists have received offer.
Muzis
Two-days hakaton to develop AI-solution for music streaming service.

Result: 23 participant teams, (offline + online), 3 data scientist received offers
Developed solutions
ABInBev
Developed the algorithm to recognise company's products in photos
2017
ABInBev
Digital Assistant of the visitor of Confederations Cup
2017
Gasprom Media Holding
Deeloped recommendation algorithm for personalised music based on machine learning technologies
2017
Central election committee of Russian Federation
Prototype of voice assistant for citizens
2017
Gasprom Media Holding
Digital platform for the distribution of video content via instant messenger

2017
AstraZeneca
Automation of internal support for more than 1,000 field employees
2018
MIR payment system
Digital assistant of the national competition participants in the framework of the PR campaign of the payment system
2018
ABInBev
Digital Assistant for visitors of the World Cup
2018
Reginal govermnemt of Russian Federation
Vehicle type recognition solution for road cameras
2018
Gasprom Media Holding
Digital Assistant accompanying employees along the business process
2018
Unilever
Digital Assistant to an employee of the company (automation of internal support service)
2018
Projects in research and implementation
Large construction company
Digital industrial platform
2018
Large logistics company
Rail defect recognition automation system
2018
Large FMCG company
Algorithm of recognition of inventory numbers on the equipment in order to speed up and optimize the inventory process 10 times
Algorithm for recognition of brand products on store shelves in order to accelerate and optimize the merchandising process
2018
Large industrial holding (chemistry)
Digital Assistant for Human resources
2018
Moscow pharmacy chain
Digital Assistant for the customer
2018
Major real estate company
The system of recognition and tracking of people and streams of people
2018