Banking and Insurance

Bancassurance is an arrangement in which the insurer sells products through the bank's channels. Banks receive risk-free commission income; the risk bearer remains the insurer. Insurers are important to banks as a source of share capital and financing: they invest large sums in debt and equity markets.

Digitalisation, new forms of risk and client requirements are changing the insurance industry: lower prices and higher-quality products for consumers and a more profitable model for insurers. The future of banking with remote methods, according to the source, will be different from today: changing consumer expectations, new technologies and business models require strategies already now — for servicing in 2030.

Assessment of the attractiveness of the branch market using GIS (a vector base of remote-sensing geospace) helps banks decide whether to reduce market presence and assess accompanying risks. Financial institutions work not with «pictures» but with information: tabular and text outputs — how much to allocate, when repayment on loans, borrowings and leases.

The most advanced direction, where remote sensing is established as a reliable, legally confirmed method of providing financial services, is agriculture. Owners of small farms will need up to 180 billion dollars of working capital and loans; they make up about 31% of the world's population and often have no financial records in banks.

The market for digitalisation of RS in agriculture is divided in the source into: remote yield monitoring; crop scouting; field mapping; variable-rate application; weather tracking and forecasting; inventory management; farm labour management; financial management; other (demand forecasting, customer management, accounts payable and receivable). The yield-monitoring sector held the largest share through 2030; the largest share of the world market for digital agricultural RS is expected in the Asia-Pacific region.

Satellite data give financial institutions access to yields and remote valuation of fields — for risk assessment when lending to small farmers. Crop-monitoring platforms open historical farm data, vegetation indices and productivity without a mandatory agent visit to every field.

Insurance companies use RS to quickly understand the affected regions and forecast the estimated claim amount for properties. Spatio-temporal analysis and ML/AI on images before and after the event show the affected areas, intensity and the size of material damage.

Areas with a large number of private individual houses are often better suited to loss assessment by remote sensing. In cities high-rise buildings can interfere with flood survey; insurers can complement RS with hydrological data: satellites show river flow and the floodplain, real-time hydrology covers the satellite revisit interval at flood peak and signal interference.

Source case: floods on the east coast of Australia in 2021. Imagery was used to detect floods and the scale of losses: water depth, inundation area, duration; inland (pluvial and river) and coastal (storm surge) floods. SAR provided a three-dimensional risk assessment while the event was still developing. In New South Wales traces of the event overlaid on portfolio data showed that estimated losses were lower than the initial assessment. SAR records water bodies; DEM, if available, — inundation depth for risk-level estimates.

Alternative credit scoring: low-income people in rural areas often have no stable credit history. Satellite images provide historical data on farmland and construction sites for value assessment. Accurate RS risk assessment is one of the main tasks of banks in lending: whether the borrower will be able to repay the loan, including in earthquake, flood, fire, chemical-production and oil-refinery zones.

Source case: a deep-learning model for yield forecast in Kenya. Financial-service providers for small farmers can assess the timing and magnitude of the harvest from imagery at low marginal cost. Models of automatic interpretation of maize, beans and potatoes were built for the summer and autumn seasons. More than 20 models were tested; the best — an InceptionV3 convolutional neural network architecture and RGB channels for maize yield forecast from summer 2016 imagery.

Industrial finance: geotags of factories, warehouses and stores; periodic surveys by UAV and mobile 3D cameras of actual inventories; discrepancies with the borrower's reports — grounds for a detailed check, including before and after large payments. Infrastructure financing: 3D from satellite images of sites, supplemented by UAV surveys, — kilometres of roads, area and volume of extraction, new telecommunications and power-line towers, the number of completed housing units.

Time series of free multispectral data: 16-day composite NDVI maps at 250 m geometric resolution for 2000–2018 described medium-term yield trends at the regional and macro-agricultural level. Sentinel-2 series at 10 m resolution detailed the assessment at field level and the linking of insurance risk to local crop condition. An operational mathematical model for calibrating premiums at the year and region level was proposed; crops — wheat, maize, ryegrass and meadows, 2016 and 2017, Piedmont region (north-west Italy).

Remote sensing was used as an indicator of socio-economic features: building size, distance between houses, street width, lack of trees and greenery. Together with multi-criteria weighted GIS overlay this was applied to search for food-bank locations. The study is related to UN Sustainable Development Goals Nos. 2, 11 and 12. Source limitation: a census at neighbourhood level is more accurate, the work was done at district level; expatriates change place of residence, which complicates the GIS base of bank users.

GIS in banking: charts, maps and 3D surface models — hills, mountains, landslides, trees, buildings, streets, rivers. Visual representation shows the relationship of locations and helps choose a retail or shopping-centre site relative to other stores and the presence of customers. Source illustrations (the Banking Salesforce Customer 360 platform and case schemes) were not transferred to the gallery.