Drawing up Real time Risk Weighted T&Cs for Merchant Banks
The Cards and payments industry, by dint of sheer operational complexities and significant extraneous stimuli, has always posed special challenges for driving AI actionability.
This extreme complexity stems from the interdependence between various players in the value chain.
Attempted a rough mapping of some of the larger players and their key operations within the industry.
It makes sense to understand this complex value chain. An overly simplistic transactional flow could look like this:
One of the key problem statements I have repeatedly heard some Acquirer and Processor clients raise in the past few years is that of finding a smart, yet expeditious way of acquiring merchants, and making optimum risk-weighted offers to them. The intent is to help expand the acquirer’s market to onboard the maximum number of merchants, while ensuring ‘bad cases’ are filtered out or priced appropriately.
That would entail risk neutral pricing and T&Cs.
But first, it is important to understand the components of the T&C price. This would comprise two key components — Price for “services” and Premium for any risks undertaken by the acquirer.
- Pricing of services needs to factor in cost of services such as ‘transaction authorisation’, ‘collection of sales slips’, ‘clearing of transactions to associations or issuers’, ‘helping in settlement of transactions’, ‘reporting’ and other case specific.
- Risk premium is estimated using various metrics and models to ensure coverage of a broad spectrum of events and market triggers as depicted below:
The key unknown factor for building the personalized T&C is individual level risk scores for each new merchant.
Net risk indexes are created for each new prospect based on identified risks as well as potential of business.
The latter would include indicators such as monthly transaction amount processed, average card transaction value, industry (ex Travel industry normally witnesses higher chargebacks than stable business like FMCG), outlook of the country a business operates in, demand and growth projections for the merchant etc
Various T&C structures can be designed in real time using AI models to recommend the most optimum contract that represents adequately covered interests of both Merchant and Acquirer.
Our solution flow is depicted below:
The process runs on an AWS TPU environment and is built using some native AWS services and Python/Dash programming.
The key outcome is to generate a real time risk weighted T&C for each merchant who signs up, expediting the offer process. Additionally we linked the offer with the closest available ‘advisor’. This allows for interventions in case of counter questions by merchants.
Some of the additional questions answered through this application, for the same project, included the following:
PS: Visuals and workflows are sanitized for all proprietary data and outputs; and meant to be emblematic.
#Risk #Cards #Payments #RiskPricing #Pricing #MerchantRisk
