A summit report on Sber’s Habr blog detailed lending automation targets and AI uses in robotics, agriculture, medicine, media and construction.
Sber expects AI to automate 75% of loans to companies and 100% of loans to individuals by early 2027. These are targets, not current automation rates. The figures appeared in a first-person report published October 8 on Sber’s Habr company blog after the Moscow Startup Summit, alongside accounts of AI applications in five industry groups.
Sberbank CEO Herman Gref said a corporate loan can currently be approved in seven minutes. He also attributed Russian neural networks’ lag behind global leaders primarily to scarce, expensive computing capacity and estimated that Russia was about six months behind China.
In robotics, summit participants favored selling measurable outcomes through a “robot as a service” model rather than selling hardware alone, particularly in cleaning, warehouse logistics and delivery. The blog report said Yandex Robotics is developing subscriptions with performance guarantees, MTS Web Services is building a vendor-independent physical AI platform with a skills marketplace, and Sber and X5 are testing robots in retail spaces.
Agricultural applications pair AI with satellite imagery and field sensors to compare crop conditions with weather, disease and pest data, helping determine where water, fertilizer or treatment is needed. In media, the report said AI has made content production faster and cheaper while attention remains limited: summit speakers estimated people have four to six free hours per day. It also noted negative reactions to low-quality AI content presented as authentic.
In healthcare, a near-term use is processing screening results and sorting patients by risk, while leaving the final decision to a doctor. The report cited about 2,000 telesurgery procedures by mid-2026. It also said the Kurikan project was preparing a cross-border robotic operation on a child, connecting Shanghai and Irkutsk with signal latency of up to 120 milliseconds.
For construction, the report described AI as one part of a building-wide digital model. Proposed uses include analyzing initial design data, monitoring safety and equipment downtime with computer vision, and comparing construction sites with 3D models through laser scanning. Participants estimated that digital construction management could save 7–8% of project costs, but cited weak digital maturity, data problems, employee resistance and model hallucinations as constraints.
Across the five sectors, the common requirement is an operating environment built around the technology. Prepared data, redesigned workflows, compatible infrastructure and explicit responsibility determine whether an AI system can deliver practical value. That is particularly consequential in robotics, medicine and construction, where technical operation alone does not establish economic viability or justify autonomous decisions.
Sources
Originally published by AI Feed: Sber Outlines AI Plans for Lending and Five Industries.
Top comments (0)