AI comes into its own in the fight against financial crime

Mon­ey laun­der­ing tech­niques have evolved sig­nif­i­cant­ly as crim­i­nals have lever­aged tech­no­log­i­cal advances. With busi­ness and soci­ety becom­ing more con­nect­ed, finan­cial crim­i­nals have adapt­ed quick­ly. Mean­while, increas­ing­ly strin­gent reg­u­la­tions and increased num­bers of fines levied against banks mean finan­cial ser­vices organ­i­sa­tions are spend­ing $180.9 bil­lion annu­al­ly on finan­cial crime com­pli­ance, 62 per cent of which goes on labour expen­di­ture in the Europe, Mid­dle East and Africa region, the Lex­is­Nex­is Risk Solu­tions Glob­al Study found. All of this to recov­er less than 1 per cent of all the crim­i­nal pro­ceeds, accord­ing to the Unit­ed Nations.

Iden­ti­fy­ing this activ­i­ty using tra­di­tion­al meth­ods is extreme­ly dif­fi­cult. While banks have been seek­ing to adopt tech­nol­o­gy that does this more effi­cient­ly by sim­ply dou­bling down on exist­ing sys­tems, anti-mon­ey laun­der­ing (AML) process­es tend to remain sig­nif­i­cant­ly siloed, with a lack of cohe­sion between sys­tems and depart­ments. Lega­cy tech­nol­o­gy is a great inhibitor, slow­ing down banks at a time when open-source tech­nol­o­gy is enabling crim­i­nals to adapt and evolve more quick­ly. It’s imper­a­tive that banks find ways to look for vul­ner­a­bil­i­ties in sys­tems more gen­er­al­ly, and intel­li­gent­ly, while pro­mot­ing clos­er align­ment between depart­ments.

“Banks have always kept infor­ma­tion restrict­ed, shar­ing it on a need-to-know basis between depart­ments. Con­se­quent­ly, they’re unable to get a holis­tic view of their clients,” says Dr Janet Bas­ti­man, head of ana­lyt­ics at regtech com­pa­ny Napi­er, whose intel­li­gent com­pli­ance plat­form helps banks increase effi­cien­cy and min­imise risk. “They tend to have mul­ti­ple teams for onboard­ing, client life-cycle man­age­ment, trans­ac­tion mon­i­tor­ing and sales. Each of these teams can also be split geo­graph­i­cal­ly, so they don’t inter­act well or share data and insights. Things are quite lit­er­al­ly falling through the gaps.

“The cyber­se­cu­ri­ty sec­tor is very good at com­mu­ni­cat­ing new threats quick­ly, so every­body can imme­di­ate­ly start patch­ing. We need to have the same approach with mon­ey laun­der­ing. If some­body knows there has been sus­pi­cious activ­i­ty, that infor­ma­tion and how to recog­nise the new pat­terns needs to spread quick­ly to pre­vent recur­rence. This isn’t hap­pen­ing because banks don’t have the sys­tems, process­es or tech­nol­o­gy to share the infor­ma­tion and get a tru­ly holis­tic view.”

Napier’s award-win­ning com­pli­ance plat­form pro­vides the com­pre­hen­sive view of cus­tomers that banks need. The company’s intel­li­gent approach, which suc­cess­ful­ly com­bines big data tech­nolo­gies with arti­fi­cial intel­li­gence (AI), robot­ic process automa­tion and machine learn­ing, is applied to under­pin pol­i­cy, process and pro­ce­dure. The Napi­er plat­form is fast, scal­able and mod­u­lar, mean­ing that finan­cial insti­tu­tions don’t need to replace their exist­ing sys­tems imme­di­ate­ly and can build their sophis­ti­ca­tion incre­men­tal­ly.

The soft­ware helps dif­fer­ent depart­ments work togeth­er more effec­tive­ly. As infor­ma­tion runs through the sys­tem, it forms a top-lev­el overview that then pro­vides alerts to the appro­pri­ate teams at the right time. Client onboard­ing and KYC (know your cus­tomer) checks, for exam­ple, can be pow­ered by con­tex­tu­alised infor­ma­tion from numer­ous sources, and com­pared with behav­iour from oth­er cus­tomers and enti­ties. A customer’s behav­iour may look nor­mal when viewed in iso­la­tion, but look­ing at it more holis­ti­cal­ly, next to all oth­er sources of infor­ma­tion, could show some unusu­al pat­terns.

“We want to make com­pli­ance offi­cers sleep eas­i­ly. With Napi­er, banks can bet­ter under­stand the fun­da­men­tal inter­con­nect­ed­ness of their data,” Bas­ti­man adds. “You have the cus­tomers and how they’re con­nect­ed to oth­er cus­tomers and all their trans­ac­tions, and being able to see that spi­der-like view real­ly expos­es any incon­sis­ten­cies. But to do this you need that holis­tic view, a cus­tomer-cen­tric approach, rather than just look­ing at siloed trans­ac­tions.”

Tak­ing this more intel­li­gent approach also brings oth­er ben­e­fits, includ­ing free­ing up the time of com­pli­ance ana­lysts who no longer have to sift through reams of trans­ac­tions to try to spot pat­terns. Reduc­ing the num­ber of peo­ple required on these kinds of inves­ti­ga­tions, and feed­ing peo­ple with accu­rate infor­ma­tion quick­ly, means humans can focus on more sophis­ti­cat­ed tasks.

The tech­nol­o­gy is pow­er­ing bet­ter explain­abil­i­ty in a reg­u­la­to­ry sense too. It’s not enough to say an issue was flagged by AI. Ana­lysts need the detail in a sim­ple, digestible lan­guage so they can explain to reg­u­la­tors exact­ly what caused con­cern. His­tor­i­cal­ly, AI has not been suc­cess­ful here, with any explain­abil­i­ty focused only on met­rics for data sci­en­tists. Napier’s Client Activ­i­ty Review has AI flags that show in plain Eng­lish what the unusu­al trans­ac­tion was in that peri­od for the client, and why it was unusu­al. This enables any­one in the team to work with the insights, with­out requir­ing a data sci­en­tist to inter­pret the data.

“Crim­i­nals will always be try­ing to hide their activ­i­ties and be one step ahead,” says Luca Primer­a­no, Chief AI Offi­cer at Napi­er. “As eco­nom­ics change in the world, whether that’s polit­i­cal or through major world events we’ve seen this year, it’s going to bring them new oppor­tu­ni­ties and also new chal­lenges. The finan­cial indus­try must real­ly adapt to those chal­lenges while stop­ping the oppor­tu­ni­ties for crim­i­nals as fast as they can.

“An AI engine can look at the trans­ac­tion­al activ­i­ties of cus­tomers much bet­ter than a human. It’s a com­plete­ly inde­pen­dent set of lens­es that can go through bil­lions of trans­ac­tions across mul­ti­ple dimen­sions to detect anom­alies, some­thing that would take humans years to com­plete at great cost. The AI then col­lates that infor­ma­tion into a sim­ple sum­ma­ry of the top sus­pi­cious behav­iours, includ­ing why they are unusu­al, so that a human ana­lyst can make bet­ter deci­sions. Napi­er pro­vides a com­plete­ly uni­fied solu­tion with these capa­bil­i­ties, elim­i­nat­ing silos and com­bin­ing every­thing togeth­er to gain that holis­tic view of cus­tomer behav­iour. This helps banks work faster and smarter to fight finan­cial crime, while at the same time reduc­ing costs.”

For more infor­ma­tion, vis­it napier.ai

Pro­mot­ed by Napi­er