He explained that the reservation chart is drawn four hours before a train’s departure. After that, the AI model can optimise seats as it is able to tell x number of seats will be vacant after say two stations. So, the allocation of seats to those in the waiting list can be accordingly. “The AI model is able to study data and predict availability of seats, leading to greater confirmations,” Vaishnaw said. Traditionally, the Railway Board allocates seats per station to ensure that at least some passengers get accommodated from each stop of a train. Using AI, the railways now identifies stations where demand is higher and rejigs this allocation.
“In our railway kitchens, we have installed Al-based cameras. They are detecting abnormal behaviour and cleanliness. In a pilot project in Pune, 100% verification of washed bed sheets is being done using Al-trained cameras. Earlier, only 2% verification was possible through sample checks. Customer satisfaction has gone up 100% in this area,” the minister said.