Robotics En­abled Per­sonal Loan Un­der­writ­ing Process for a Large Re­tail Bank

Banking Frontiers - - Impact Survey -

Typ­i­cally, in Asian banks the vol­ume of new cus­tomers and new ac­counts on both li­a­bil­i­ties and as­sets side are very high. The ex­tent of pen­e­tra­tion of the in­ter­net, dig­i­tal lit­er­acy among ru­ral and semi-ur­ban ter­rains, speed of pro­cess­ing are some of the fac­tors which con­strains Asian in­dus­try to man­age a med­ley of dig­i­tal and phys­i­cal ap­pli­ca­tion forms for data col­lec­tion. High vol­umes cou­pled with the ever de­mand­ing turn around times (TAT) ow­ing to com­pe­ti­tion and re­newed fo­cus on cost re­duc­tion paves the way for im­ple­men­ta­tion of new age so­lu­tions. For one of our clients, a large pri­vate sec­tor bank, we have de­vised a Robotics Loan Process So­lu­tion that has three dis­tinct fea­tures, which en­ables near 100% au­to­ma­tion and hence all the as­so­ci­ated ben­e­fits of cost op­ti­miza­tion, im­proved TAT and lower risk of fraud­u­lent pro­cess­ing of in­for­ma­tion are passed on to client.

Fea­ture 1: Phys­i­cal form con­verted into dig­i­tal data by us­ing ICR/OCR tech­nol­ogy. As ICR/OCR has lim­ited suc­cess rate in case of man­ual script, to en­hance ac­cu­racy we have in­tro­duced what we call a ‘Robotics Process en­gine’. The ex­tracted data us­ing an IC/OCR en­gine passes through these au­to­mated rules; a ‘ro­bot’ val­i­dates the data col­lected and al­ters val­ues based on a pre-pro­cessed data­base. The data­base will be en­riched with ev­ery cor­rec­tion that is made and has the abil­ity to store and re­trieve millions of pat­terns/com­bi­na­tions. Ro­botic process au­to­ma­tion (RPA) is the ap­pli­ca­tion of tech­nol­ogy that al­lows to con­fig­ure soft­ware or a ‘ro­bot’ to cap­ture and in­ter­pret ex­ist­ing ap­pli­ca­tions for pro­cess­ing a trans­ac­tion, ma­nip­u­lat­ing data, trig­ger­ing re­sponses and com­mu­ni­cat­ing with other dig­i­tal sys­tems.

Fea­ture 2: Writer In­for­ma­tion’s Loan Pro­cess­ing So­lu­tion has been cre­ated with the flex­i­bil­ity of cap­tur­ing KYC in­for­ma­tion from mul­ti­ple doc­u­ment types, namely the Pass­port, PAN Card, Aad­haar

Card and util­ity bills along with a dig­i­tized copy of the KYC doc­u­ment. The data cap­ture phase acts as the ini­tial in­ward­ing phase of an en­rol­ment process and in­te­grates with a de­ci­sion en­gine. The de­ci­sion en­gine has been built us­ing an­a­lyt­i­cal mod­els to de­rive in­fer­ences of data cap­tured and cross check the ve­rac­ity of the data read by the ap­pli­ca­tion au­to­mat­i­cally. Checks and bal­ances en­sure there is an ad­di­tional layer of data val­i­da­tion.

Fea­ture 3: This pro­cessed data helps to un­der­write the credit score and then pass the trans­ac­tion to the Core Busi­ness Ap­pli­ca­tion for credit de­ci­sion. Credit de­ci­sion in­volves a del­e­ga­tion of pow­ers and hence has to be routed to dif­fer­ent stake­hold­ers for ap­proval. To fa­cil­i­tate auto-rout­ing of the pro­pos­als, a work-flow tool was in­te­grated with de­ci­sion en­gine and this makes Writer In­for­ma­tion’s so­lu­tion fully in­te­grated and au­to­mated.

The other high­light of the Writer Info-rma­tion’s Loan Process So­lu­tion is that this can work in a ‘tab­based’ Data Cap­ture model for sin­gle case han­dling as well as in a server-model for bulk pro­cesses across thou­sands of records us­ing the robotics so­lu­tion.

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.