Nascent AI head­ing to­wards Cog­ni­fi­ca­tion

Banks and fi­nan­cial in­sti­tu­tions speak of AI as the next fron­tier but at present its uses are just pe­riph­eral though the scope is in­fi­nite

Banking Frontiers - - News - Mo­han@bank­ingfron­tiers.com

While AI is in­creas­ingly be­com­ing the next-in-line fad for the bank­ing com­mu­nity, right now its uses are con­fined to chat­bots, IPAs and sim­i­lar in­ter­ac­tive tools. What ex­perts feel is that un­less NLP, or Nat­u­ral Lan­guage Pro­cess­ing, is per­fected and de­ployed, there will be only pe­riph­eral use of AI in the way banks in­ter­act with its cus­tomers.

It is per­ceived that AI will have ap­pli­ca­tions and bet­ter uses in ar­eas like per­sonal fi­nan­cial man­age­ment, ad­vi­sory ser­vices, ad­ver­tis­ing, rec­om­men­da­tions to cus­tomers and se­cu­rity sys­tems. Ka­sisto, the US firm that de­vel­ops con­ver­sa­tion plat­forms for banks to add vir­tual per­sonal as­sis­tants to their mo­bile and tablet of­fer­ings, has said that it is de­vel­op­ing an om­nichan­nel prod­uct that will ef­fec­tively func­tion as a ‘bank­ing brain’ that can have con­tex­tual con­ver­sa­tions with cus­tomers based on what it knows about the cus­tomers, mak­ing use of cus­tomer data to cre­ate more per­son­al­ized con­ver­sa­tions. The think­ing is shift­ing to tech­nol­ogy that can be ap­plied to a va­ri­ety of chan­nels. Master­card and Stan­dard Char­tered Bank are the ini­tial users of this prod­uct.

RPA

There is this new process of Ro­botic Process Au­to­ma­tion, or RPA, which global banks are in­creas­ingly us­ing re­cently, mainly to re­duce costs and move from ser­vices through la­bor to ser­vices through soft­ware. It in­volves the cre­ation of vir­tual work­force, which has helped these banks to some­times even elim­i­nate hu­man in­ter­ven­tion in de­ci­sion­mak­ing, in ex­e­cu­tion. Some banks in the west, which have been us­ing the ser­vices of lawyers to fi­nal­ize loan agree­ments, are now us­ing AI sys­tems, which have not just cut down costs, but have also re­duced the level of er­rors re­sult­ing from hu­man judg­ment.

The pre­dic­tion is that AI will be able to in­de­pen­dently an­a­lyze what is there in the whole of dig­i­tal world, mean­ing the in­ter­net, ap­ply this to in­ter­nal data and cre­ate in­tel­li­gence based on which so­lu­tions are cre­ated.

DEF­I­NITE PUSH

“Adop­tion of AI is no longer an op­tion or a show­case item. Ma­jor banks are al­ready

ex­per­i­ment­ing with AI so­lu­tions in some way or the other and are now solv­ing real prob­lems with it,” says Anup Purohit, CIO, Busi­ness & Dig­i­tal Tech­nol­ogy So­lu­tions at YES Bank. “Early ver­sions of chat­bots ser­viced cus­tomer FAQs. How­ever, more and more banks are now adding main­stream bank­ing ser­vices into chat­bots and are in­te­grat­ing their AI so­lu­tions into var­i­ous chat ap­pli­ca­tions. There is a def­i­nite push by banks to in­crease the us­age of their AI so­lu­tions among cus­tomers. Banks are also us­ing AI as en­abler for their em­ploy­ees. They are em­pow­er­ing their em­ploy­ees with chat­bots that an­swer process and prod­uct spe­cific queries thereby bring­ing down pro­cess­ing time and in­creas­ing cus­tomer de­light. AI has al­ready made in­roads into the BI and cy­ber se­cu­rity do­mains. There are live ex­am­ples of ML so­lu­tions an­a­lyz­ing net­work traf­fic to iden­tify po­ten­tial rogue con­nec­tions,” he says.

San­jay Sharma, head - Tech­nol­ogy, In­no­va­tion & Cus­tomer Ful­fil­ment at RBL Bank, be­lieves that each bank has had a unique jour­ney in AI, de­pend­ing upon its needs and growth strat­egy. How­ever, the key ap­pli­ca­tion has been cus­tomer fac­ing chat­bots, an­swer­ing FAQs from cus­tomers, con­duct­ing ba­sic cus­tomer trans­ac­tions, al­low­ing banks to ra­tio­nal­ize costs and to help its staff de­vote more time for value added work. “A few banks have been ex­per­i­ment­ing with other AI driven cus­tomer fac­ing ap­pli­ca­tions such as as­sess­ing cus­tomer feed­back in real time, of­fer­ing deep cus­tomiza­tions based on cus­tomer in­sights gen­er­ated across so­cial me­dia. There have been a few in­stances where banks have tried to con­duct pi­lots in in­tel­li­gent back end au­to­ma­tion as well, as a next step of Ro­botic Process Au­to­ma­tion,” says he.

He also main­tains that banks are cur­rently test­ing many AI driven ap­pli­ca­tions, with adop­tion of a few for main­stream con­sump­tion. “The year 2018 could be a year where AI ex­per­i­ments in fin­tech/ in­no­va­tion labs of banks would be con­sumed in the banks’ busi­nesses on a mass scale. To con­tin­u­ously en­hance AI ca­pa­bil­i­ties and help in achiev­ing the growth, it is nec­es­sary for In­dian banks to de­velop in­ter­nal AI ca­pa­bil­i­ties and for­mu­late a wellde­fined AI strat­egy,” he adds.

NASCENT STAGE

How­ever, Kal­pana B, part­ner and head­Robotics and Cog­ni­tive Au­to­ma­tion, KPMG in In­dia, does not think banks and fi­nan­cial in­sti­tu­tions in In­dia are mak­ing use of AI in a worth­while man­ner. Says she: “Well frankly, I am not aware of any AI ap­pli­ca­tions in use by In­dian banks. I am not dis­re­gard­ing the con­ver­sa­tional chat bots, but these are still at a very nascent stage. The con­ver­sa­tional chat­bots are yet to be ad­vanced to a de­gree where it can sup­port be­yond pre­de­fined queries.”

She says cur­rently the adop­tion of AI in bank­ing is at its nascent stage. “To be hon­est, we are in need of use­ful data in or­der to train any­thing in in­tel­li­gence. I see a rapid catch up in the next three years to avoid ex­is­ten­tial threats,” she adds.

Ac­cord­ing to Ra­jiv Ku­mar, MD & CEO, Uni­ver­sal Sompo Gen­eral In­surance, a host of AI-re­lated tech­nolo­gies such as NLP, Com­puter Vi­sion, robotics, ma­chine learn­ing and speech recog­ni­tion have sur­faced in the in­dus­try and have sub­stan­tially pro­gressed over the years to co­a­lesce into In­dian sys­tems that are ca­pa­ble to per­form ac­tiv­i­ties such as think, learn and con­tin­u­ously adapt. “How­ever, at present, the eas­ily-con­fig­urable and im­ple­mentable so­lu­tions us­ing AI, such as chat­bots, per­sonal fi­nance track­ers and ad­vi­sors with the rule-based con­fig­u­ra­tion to re­solve pol­i­cy­hold­ers’ queries are in use and are also cus­tomer faced in the in­surance seg­ment,” says he.

MAIN­STREAM USE

AI is gain­ing on its ca­pa­bil­i­ties at a fast pace fa­cil­i­tated by bet­ter com­put­ing power and high cal­iber hard­ware. Along­side, ro­bots are gain­ing foothold, re­plac­ing hu­man la­bor in crit­i­cal ar­eas. Chat­bots and IPAs are not the real do­main power of AI. Purohit says AI is de­fined as the abil­ity of a com­puter/soft­ware pro­gram to per­form tasks nor­mally re­quir­ing hu­man in­tel­li­gence. “Hu­man in­tel­li­gence re­lies on past data/mem­ory, cur­rent data points and at times im­pact of a de­ci­sion on the fore­see­able fu­ture to pro­vide the rel­e­vant an­swer, the best-al­ter­na­tive or a de­ci­sion. AI so­lu­tions are in­creas­ingly be­com­ing bet­ter at mim­ick­ing this hu­man be­hav­ior at far greater speeds as com­pared to hu­mans. Be­sides, they do not have mem­ory lim­i­ta­tions like hu­mans. This would be the main­stream use for AI – aug­ment hu­man de­ci­sions with ‘best-al­ter­na­tive’ data points and go­ing a step ahead, de­cide on be­half of hu­mans,” says he.

He adds that while AI as a plat­form can cater to any kind of in­dus­try, fin­techs are tak­ing the effort to de­velop the li­braries to make their AI so­lu­tions bank­ing-aware. They are also cre­at­ing li­braries spe­cific to cer­tain mar­kets/re­gion/coun­try. This al­lows for faster adop­tion of the AI so­lu­tion and avoids dis­cov­ery of mis­takes over longer pe­ri­ods of time.

IN­NO­VA­TIVE USES

“Banks too are get­ting in­no­va­tive and im­ple­ment­ing AI so­lu­tions in a novel way. For ex­am­ple, use of vi­sion based AI so­lu­tions can an­a­lyze video feeds from ATMs in real-time. A per­son stand­ing too close to the ATM, bend­ing down too much or mak­ing un­usual move­ments can be tagged as sus­pi­cious. They can trig­ger an alarm or dis­able the ATM till the ser­vice provider scans it for snoop­ing de­vices or pil­fer­age,” says Purohit.

Sharma is of the view that banks

need to ex­pand bound­aries of AI be­yond con­ver­sa­tional an­a­lyt­ics. He says a few fin­techs are of­fer­ing ap­pli­ca­tions fo­cused on CXO dash­board for in­ter­nal con­sump­tion, which can be lever­aged. A chat­bot can an­swer a few crit­i­cal ques­tions such as which branch is likely to achieve max­i­mum growth in terms of new busi­ness in the next quar­ter and which branch is go­ing to lag; what are the skills, which are in the high­est de­mand in the mar­ket; how have the last 5 gen­eral elec­tions im­pacted bank­ing; are there any mea­sures suc­cess­fully adopted by banks in other re­gions: and what is the % adop­tion of mo­bile pay­ments in tier II cities and be­yond.

“There are a few in­no­va­tive uses of AI in fields such as fraud de­tec­tion, AML, com­pli­ance, au­to­mated au­dits, etc. which hold good prom­ise,” he adds.

For Dr N. Ra­jen­dran, CTO, Na­tional Pay­ments Cor­po­ra­tion of In­dia, while AI will find ap­pli­ca­tions in busi­ness an­a­lyt­ics and fraud de­tec­tion sys­tems, it will be real time an­a­lyt­ics with AI and ML tools that would im­prove the busi­ness de­ci­sions, fraud pre­ven­tion, cy­ber se­cu­rity and of pre­dic­tion in these ar­eas.

Ra­jiv Ku­mar says ap­pli­ca­tions such as wear­able de­vices, telem­at­ics and chat­bots ca­pa­ble of query re­solv­ing hold prom­ise for the in­surance sec­tor. “As you know, IRDAI has rec­og­nized the role of well­ness as­pects in risk as­sess­ment and prod­uct de­sign for health in­surance of­fer­ings and use of telem­at­ics for the mo­tor seg­ment. The in­surance seg­ment is grad­u­ally evolv­ing on these fronts,” says he,

COG­NI­FI­CA­TION

The pre­dic­tion is that the next stage of dig­i­ti­za­tion and au­to­ma­tion will be the process of ‘cog­ni­fi­ca­tion’. It is de­scribed as the com­bi­na­tion of IoT and AI. It is be­lieved that the syn­er­gis­tic ef­fect of ap­ply­ing AI to con­nected de­vices will be in­cal­cu­la­ble. Purohit be­lieves this will be the nat­u­ral pro­gres­sion. “Cur­rently AI sys­tems ex­posed to users are man­ag­ing low risk queries or pro­cesses. As the con­fi­dence in these sys­tem in­creases with more us­age and more learn­ing data, users will com­pare their de­ci­sion with the one pro­vided by the AI sys­tem. At some point, we will al­low the AI sys­tem to make these de­ci­sions,” says he.

The best ex­am­ple would be the traf­fic nav­i­ga­tion ap­pli­ca­tions. A large per­cent­age of the users com­pletely go by the ‘short­est travel time’ path pro­vided by the ap­pli­ca­tion, he ar­gues.

Sharma says a few spe­cific ac­tion­able items to achieve cog­ni­fi­ca­tion are:

Se­lect the next best ac­tion

Achieve N=1 level cus­tomiza­tion

“AI will help an­a­lyze data across mul­ti­ple un­struc­tured, semi struc­tured and struc­tured sources and gen­er­ate mean­ing­ful and ac­tion­able in­sights to of­fer cus­tom­ized and dif­fer­en­ti­ated so­lu­tions. This should help a bank to be­come more cog­ni­tive - a bank that thinks and in­sights driven over the next few years,” says he.

Kal­pana too be­lieves that AI will be used with pro­cesses to make spe­cific de­ci­sions. “I think it is very much pos­si­ble. There is a con­tin­u­ous evo­lu­tion that is tak­ing place in the tech­no­log­i­cal do­main and AI has def­i­nitely made its way from sci-fi movies into our daily life de­ci­sion-mak­ing,” she says.

Ra­jiv Ku­mar ex­plains that cog­ni­fi­ca­tion in­volves ad­vanced tech­nol­ogy plat­forms that can ad­dress com­plex sit­u­a­tions that are char­ac­ter­ized by am­bi­gu­ity and un­cer­tainty. He has words of cau­tion: “Cog­ni­tive com­put­ing has be­gun to aug­ment busi­ness de­ci­sions and power per­for­mance right along­side hu­man thought process and tra­di­tional an­a­lyt­ics. How­ever, there are a good num­ber of in­dus­try lead­ers who would want to see cog­ni­fi­ca­tion in the area of per­for­mance man­age­ment but for me, the role of hu­man de­ci­sion should al­ways be val­ued more than machines and there should be some guide­lines or up­per lim­its to use AI.”

FIN­TECHS AND AI

Purohit points out that banks to­day are in­creas­ingly part­ner­ing with fin­techs through their co-in­no­va­tion labs. “Fin­techs, by their very na­ture, fo­cus on solv­ing a spe­cific prob­lem and/or cre­at­ing a spe­cific prod­uct within short times­pans,” says he. “They mostly tend to use the lat­est tech­nolo­gies and are there­fore adopt­ing AI in a big way in their so­lu­tions. Which means banks now have easy ac­cess to the lat­est AI so­lu­tions along with the nec­es­sary skillsets and knowl­edge­base to im­ple­ment it.”

He goes on to add: “It’s a win-win sit­u­a­tion for both. Banks get to im­ple­ment and use the lat­est AI so­lu­tions in short times­pans while fin­techs get to fine­tune their prod­uct by an­a­lyz­ing the huge datasets avail­able with the bank and feed­back from do­main ex­perts in the bank.”

Kal­pana ex­plains that fin­techs are con­tained mi­cro­cosms to start change in the sec­tor. They nor­mally cater to a spe­cific need or use and be­cause of their ex­treme fo­cus, they help banks, says she, adding: “For a va­ri­ety of tasks, it’s like or­ches­trat­ing fin­techs across a process /pro­cesses and no one has yet com­pleted the jour­ney. There is a huge dif­fer­ence if you are an in­di­vid­ual vi­o­lin/cello player ver­sus if you are one amongst many in a phil­har­monic orches­tra. Some­times we ob­serve the size of or­ga­ni­za­tions in­hibits speed and progress.”

Sharma says fin­techs and chal­lenger banks have taken a lead in de­vel­op­ing and con­sum­ing AI driven ap­pli­ca­tions and banks have been ex­per­i­ment­ing with such ap­pli­ca­tions through hackathons and ac­cel­er­a­tor pro­grams. “I see banks and fin­techs col­lab­o­rat­ing over APIs, sand­box for de­vel­op­ing in­no­va­tive AI driven so­lu­tions,” he says.

Dr Ra­jen­dran too says fin­techs and

chal­lenger banks have taken a lead in de­vel­op­ing and con­sum­ing AI driven ap­pli­ca­tions. He also men­tions that in the days to come adop­tion of AI will in­crease in the bank­ing and fi­nan­cial ser­vices sec­tor.

RE­PLAC­ING HU­MANS

Is there a prospect of the use of AI leading to re­place­ment of hu­mans in cer­tain op­er­a­tions? Or does it serve the role of ‘aug­ment­ing’ the op­er­a­tions?

Says Kal­pana: “Yes, there are pos­si­bil­i­ties where AI may re­place hu­mans if that serves the ob­jec­tives bet­ter. And yes, it would aug­ment the hu­man ca­pa­bil­ity in some other con­text. For ex­am­ple, a con­tact cen­ter in­ter­face could be elim­i­nated if AI can pre­dict what may hap­pen with a cus­tomer and pre-solve their po­ten­tial cause to call the cen­ter. At the same time, it can also aug­ment a per­sonal re­la­tion­ship man­ager to bet­ter serve his clien­tele by un­der­stand­ing their life stage /cre­ate and re­view com­plex al­gos and sug­gest what in­vest­ment best suits.”

Purohit too says AI will aug­ment and will def­i­nitely help free up hu­man band­width in cer­tain op­er­a­tions. And this freed-up band­width will be re­de­ployed to­wards new prod­uct lines/ser­vices and reach out to a wider cus­tomer base, he adds.

Stat­ing that there has been a lot of talk and de­bate around this topic, Sharma ar­gues that AI should be viewed as a mech­a­nism to help a bank be­come more in­tel­li­gent and smart. “This is pos­si­ble when em­ploy­ees can fo­cus more on value added tasks, while AI based ap­pli­ca­tions help to au­to­mate repet­i­tive, time con­sum­ing tasks. I think at least for the next few years as banks will aug­ment their AI ca­pa­bil­i­ties, AI will act as the em­ployee’s com­pan­ion to gen­er­ate re­sults bet­ter and faster and de­liver bet­ter.”

Dr Ra­jen­dran says AI will be used for au­to­ma­tion in cer­tain op­er­a­tions and it may re­duce the hu­mans in cer­tain op­er­a­tions.

HU­MAN TOUCH

In this con­text, will it be a chal­lenge for banks, in­creas­ingly us­ing AI, to re­tain the hu­man touch?

“This ques­tion has come up in the past,” says Purohit, “when banks launched ATMs, in­ter­net bank­ing, IVR bank­ing, SMS bank­ing, mo­bile bank­ing. While they re­duced or elim­i­nated the cus­tomers’ branch vis­its, these op­tions pro­vided con­ve­nience and time flex­i­bil­ity to the cus­tomers. The AI op­tion is no dif­fer­ent. The ben­e­fit it will bring in will far out­weigh the loss of hu­man touch, if any.

“I think hu­mans would do what hu­mans do best – con­nect with oth­ers,” opines Kal­pana. “I am not see­ing last mile adop­tions even in the west­ern world yet. We need to be mind­ful of fu­ture pos­si­bil­i­ties and pre­pare rather than fear on how to re­tain con­trol. Be­cause, when we fear we freeze and that may be dan­ger­ous for the or­ga­ni­za­tion.”

Sharma as­serts that the role of AI is not to re­place hu­mans but to aug­ment their ca­pa­bil­i­ties. Most banks are ex­pected to adopt a hy­brid ap­proach, he says adding “AI driven ro­bots and per­sonal re­la­tion­ship man­agers would co-ex­ist to sup­port the cus­tomers.”

Ra­jiv Ku­mar says since in­surance is driven by data, use of AI will have a huge ef­fect on the com­pany’s bot­tom line and the sat­is­fac­tion of the cus­tomer. “How­ever, in ad­vanced na­tions, AI has rev­o­lu­tion­ized the way in­sur­ers gain in­for­ma­tion from their cus­tomers but from an In­dian per­spec­tive, the con­sumer is not that much aware of new tech­nolo­gies. Hence, for an in­surance com­pany, while there is grow­ing de­mand from mil­len­ni­als for tech­no­log­i­cal ad­vance­ment, our ru­ral, semi-ur­ban cus­tomers, se­nior cit­i­zens still re­quire the hu­man touch. So, this will be go­ing to be chal­leng­ing for us, as, while on the one hand we have to in­crease our in­vest­ment for tech­no­log­i­cal de­vel­op­ments while on the other, we still have to re­tain our tra­di­tional meth­ods.”

He says few large in­surance com­pa­nies in In­dia have ini­ti­ated to uti­lize AI for ei­ther sales or cus­tomer sup­port. “In my view, the op­er­at­ing and busi­ness mod­els of in­sur­ers are evolv­ing. Tech­nol­ogy trends such as AI, ma­chine learn­ing, blockchains and ro­botic process au­to­ma­tion have sig­nif­i­cant po­ten­tial to stream­line in­surance op­er­a­tions and en­hance cus­tomer ex­pe­ri­ence,” says he.

POS­SI­BLE AP­PLI­CA­TIONS

Sharma says go­ing for­ward, AI can of­fer con­ver­sa­tional an­a­lyt­ics as one of the many streams. “With in­creas­ing in­stances of cy­ber frauds, AI could be sig­nif­i­cantly de­ployed for as­sess­ing sus­pi­cious trans­ac­tions, pat­tern recog­ni­tion and pro­tect the con­fi­den­tial cus­tomer data at an early stage. Other main­stream uses of AI could be seam­less trans­ac­tion flow, bet­ter KYC, KYB and as­sess­ing the cus­tomers with high­est prob­a­bil­ity of sales con­ver­sion.”

Kal­pana sees in AI the po­ten­tial of vir­tu­ally chang­ing the bank­ing model and dis­rupt­ing the way tra­di­tional risk and re­ward ma­tri­ces work in banks. “In my view, if we imag­ine the cur­rent back and look for AI cases, we may not do jus­tice to the art of the pos­si­ble. The key there­fore is to reimag­ine the fu­ture,” says she.

Dr Ra­jen­dran says AI is used for var­i­ous re­quire­ments of data an­a­lyt­ics and for act­ing as dig­i­tal in­ter­face for in­ter­ac­tion with var­i­ous sys­tems and end users.

Ra­jiv Ku­mar fore­sees the evo­lu­tion of en­ter­prise AI so­lu­tions, which can en­hance op­er­a­tional ef­fi­ciency, im­prove time-to-mar­ket ca­pa­bil­i­ties, en­able a more in­tel­li­gent way to sell and ser­vice cus­tomer and more. “How­ever, from an in­surance per­spec­tive, the role of AI would be in cus­tomer ser­vic­ing, un­der­writ­ing and set­tling claims,” says he.

SPE­CIFIC AR­EAS

Specif­i­cally, what about the pos­si­ble ap­pli­ca­tion of AI in ar­eas like AML, al­go­rith­mic trad­ing, fraud de­tec­tion, cus­tomer en­gage­ment, risk man­age­ment?

Says Kal­pana: “Well, AI can be used across most of the ar­eas that you’ve listed out. I can give a few ex­am­ples to show how far they have been ap­plied in a few of the ar­eas. In the area of Algo trad­ing, AI is used for de­vel­op­ing trad­ing strategies and sug­gest­ing port­fo­lios to clients. AI in fraud de­tec­tion is seen in the form of re­duced time to scru­ti­nize ev­ery fi­nan­cial trans­ac­tion. Cus­tomer en­gage­ment is en­hanced thor­oughly with the ap­pli­ca­tion of AI, by mak­ing it su­perbly in­ter­ac­tive and per­son­al­ized. I sus­pect that AML screen­ing would be quite dif­fer­ent if AI is ap­plied and it has to work in con­sor­tium across banks and reg­u­la­tors and not just one.

Sharma is of the view that pri­or­ity for banks will be to pro­vide a more se­cure trans­act­ing en­vi­ron­ment for its users. “For the same, we can ex­pect AI to be de­ployed for: as­sess­ing the real owner of the trans­ac­tion by anal­y­sis of the trans­ac­tion sources; Know your Busi­ness (KYB) to de­tect sources of in­come, na­ture of trans­ac­tions for a strong AML and fraud risk man­age­ment plat­form; ve­loc­ity checks, tak­ing ac­tion on its own when anom­alies are de­tected; au­to­mated au­dits to dig out gen­er­ally un­fore­seen pat­terns; ac­tively hunt for do­mes­tic and global reg­u­la­tions to en­sure 100% com­pli­ance; de­liver in­sights, mar­ket news on fin­ger­tips; sig­nif­i­cantly drive cus­tomer loy­alty through sen­ti­ment anal­y­sis - as­sess­ing the mood, tone, emo­tions of the cus­tomer through nat­u­ral lan­guage pro­cess­ing ca­pa­bil­i­ties for of­fer­ing deeply cus­tom­ized so­lu­tions for that par­tic­u­lar con­text.”

AI & CBS

Has AI got the scope to be an in­te­gral part of CBS in banks?

Says Purohit: “Rapid de­vel­op­ments in AI space will make it very easy for CBS sys­tem providers to bun­dle AI into their prod­ucts. Many will in­vest in de­vel­op­ing their own AI so­lu­tions. How­ever, I fore­see CBS play­ers pro­vid­ing ca­pa­bil­i­ties to in­te­grate with other AI so­lu­tions rather than bundling one. Banks have al­ready started their AI jour­ney and with ev­ery pass­ing day, the AI so­lu­tion is get­ting bet­ter and speaking with more and more sys­tems in the bank. Switch­ing over to a bun­dled so­lu­tion will be time con­sum­ing and may not be worth the effort.”

Kal­pana up­holds the dic­tum that for any in­vest­ment, there should be a why. “I would think of it in a dif­fer­ent way. Why would you get in the way of a legacy Core Bank­ing So­lu­tion, when you can make an in­de­pen­dent AI-run dig­i­tal plat­form al­to­gether, which can take things smoothly? Even­tu­ally, the CBS can be re­placed with an AI dig­i­tal setup. Imag­ine the pre-CBS era and why we got there. Again, reimag­ine than ren­o­vate the ex­ist­ing.”

Sharma thinks over a short term, banks may fo­cus on cus­tomer fac­ing and se­cu­rity re­lated ap­pli­ca­tions for AI de­ploy­ments. How­ever, he is cer­tain CBS is one area which is ripe for in­no­va­tion. Al­ready, banks are ex­per­i­ment­ing with cloud based core of the fu­ture with open stan­dards of cod­ing, sim­pler stack, etc to en­hance per­for­mance.

AN CAIO?

Would banks ap­point chief AI officers?

Purohit says while such a ti­tle may not have been given to any­one now, banks al­ready have a per­son or a team per­form­ing this role. “I be­lieve banks have un­der­stood the need to have an en­ter­prise wide AI so­lu­tion and are in their dif­fer­ent stages of im­ple­men­ta­tion. It’s only a mat­ter of time be­fore a ti­tle is an­nounced of­fi­cially,” says he.

Sharma feels the nomen­cla­tures and ti­tles may vary with each or­ga­ni­za­tion, but the role would def­i­nitely re­quire a strong un­der­stand­ing of the uses of AI for all in­ter­nal and ex­ter­nal cus­tomer fac­ing ap­pli­ca­tions. “One im­por­tant aspect to watch out for is the level of such an of­fi­cer in the or­ga­ni­za­tion - whether he/she is a mem­ber of tech­nol­ogy trans­for­ma­tion team or rolls up di­rectly to busi­ness head or to the CEO & MD of the firm. This will in­di­cate the strat­egy which the bank has adopted with re­gards to con­sump­tion of AI,” he says.

Kal­pana has the last word: “I think we are again mak­ing the mis­take of imag­in­ing what the roles on ex­ist­ing struc­ture are. We would have CEOs to the last em­ployee in hi­er­ar­chy uti­liz­ing this tech­nol­ogy as ap­pro­pri­ate. We don’t have a chief MS ex­cel of­fi­cer, in jest, so why only for AI. We need a col­lab­o­ra­tion of di­verse skills and my view is by putting one per­son in charge, we would run risks.”

Anup Purohit be­lieves main­stream use of AI would be to aug­ment hu­man de­ci­sions with best-al­ter­na­tive data points

San­jay Sharma points out that fin­techs and chal­lenger banks have taken a lead in de­vel­op­ing and con­sum­ing AI driven ap­pli­ca­tions

Kal­pana B be­lieves use of AI is in a nascent stage and there is need of use­ful data in or­der to train any­thing in in­tel­li­gence

Ra­jiv Ku­mar fore­sees the evo­lu­tion of en­ter­prise AI so­lu­tions, which can en­hance op­er­a­tional ef­fi­ciency, im­prove time-to-mar­ket ca­pa­bil­i­ties

Dr N Ra­jen­dran feels it will be real time an­a­lyt­ics with AI and ML tools that would im­prove busi­ness de­ci­sions, fraud pre­ven­tion, cy­ber se­cu­rity

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.