Ukuthuthukiswa Komkhiqizo & QA
Ibhange
Iklayenti yiqembu lamabhange elivelele e-UAE elasungulwa ngo-1985, elaziwa ngamasevisi alo ahlukahlukene kanye nemikhiqizo: amabhange azimele, amabhange okuhweba, ukuphatha ingcebo, ukubhanga kwezinkampani, ukuhwebelana kwamanye amazwe, ezezimali zephrojekthi, ukuphathwa kwezakhiwo, nokuningi.
Ngesifiso sokuba yibhange elingu-#1 e-UAE, iklayenti, elinenani labasebenzi abangu-5,000, linikeza imikhiqizo yalo namasevisi kumakhasimende angaphezu kwesigidi esi-1.
Iklayenti lisebenzise I-Interactive Voice Response (IVR) kanye nendlela yocingo yomenzeli ukubhekana nemibuzo yamakhasimende nokwenza imisebenzi. Uhlelo olukhona beludla isikhathi kumakhasimende nama-ejenti futhi lwabeka inselele enkulu ngokombono wezenzo zokukhwabanisa ezingase zibe khona.
Ukuze bagweme le nselele, banqume ukusebenzisa isistimu esebenza ngezwi ukuze kuthuthukiswe ukuvikeleka kanye nokuqinisekisa ukuthi ungubani phakathi nezingcingo zamakhasimende kanye nemisebenzi.
Uma ubunikazi balowo ofonayo sebuqinisekisiwe, ofonayo angakwazi ukufinyelela izinsiza zebhange eziningi, njengokwengeza abazuzayo, ukudluliselwa kwemali, nokuningi - konke ngemiyalo yezwi.
Ngokusekelwe ezingxoxweni phakathi kweMaruti Techlabs neklayenti, womabili amaqembu ehlise ububanzi, obuhlanganisa okulandelayo -
Kuwumkhuba ojwayelekile ukuthi amabhange aqhube inqubo yawo yokukhuthala kwamakhasimende (KYC) ngaphambi kokunikeza amasevisi. Inqubo ye-KYC ngokuyinhloko ihlanganisa ukuqinisekiswa kwedokhumenti, njengezikweletu zezinsiza, ukuqinisekiswa kwekheli, nokuqinisekiswa kwe-biometric. Ngakho-ke, kulula kuma-ejenti ukuthi afunde imininingwane yocingo lombuzo noma isicelo.
Kodwa-ke, kuyinselele kuma-ejenti ukucacisa ukuthi ungubani umuntu okhulumayo ngakolunye uhlangothi lwekholi. Ngenxa yalesi simo esibucayi, inqubo yokuqinisekisa ikhasimende idle isikhathi sabasebenzeli ngaphandle kokunikeza uhlelo olungenangqondo lokuhlonza abakhwabanisi kumakhasimende angempela.
Abakhwabanisi namuhla banamaqhinga amaningi okufunda ukuthi ikhasimende lingubani kanye nolwazi oluyimfihlo. Iklayenti belifuna ukwethula ukubonwa kwezwi okuzenzakalelayo ukuze linikeze izindlela eziphephile nezisheshayo zokwenza imisebenzi ehlobene nebhange.
Imigomo ebalulekile bekuwukunqoba izinselele ezinjengokunciphisa isikhathi sokushintsha (okungukuthi, imizuzwana engama-45-60/ikhasimende) sokuqinisekisa amakhasimende ezingcingweni, ukunikeza ukuqinisekiswa kwezwi okungenamsebenzi esikhundleni sokuqinisekisa i-PIN, kanye nokunciphisa ubungozi bokukhwabanisa okuvela kulwazi lwekhasimende olusengozini ngokusebenzisa ubunjiniyela bomphakathi kanye ubugebengu bokweba imininingwane ebucayi.
I-Passive voice biometrics ithuthukisa ubuchwepheshe bokubona izwi ukuze ikhombe abantu ngabanye ngokusekelwe kuzimpawu zabo zephimbo ezihlukile lapho bekhuluma bengenzi lutho noma ngokwemvelo. Ukuze kuzuzwe lokhu, ithimba lethu liklame ingosi yewebhu ezofuna imvume kumakhasimende ukuze yenze iphasikhodi, ibhalise izwi ku-IVR, futhi ibhalise ikhasimende uma isigxivizo salo sezwi senziwe.
Sidizayine isixazululo esinezingxenye ezimbili esifaka iphothali yokuphatha izigxivizo zezwi kanye nohlelo lokusebenza lokuthola ukukhwabanisa. Nasi isifinyezo esifushane sesixazululo esisetshenzisiwe -
1. Ukubhaliswa kwe-Biometric Yezwi
Izindlebe zomuntu azikwazi ukubona izici ezicashile ezisiza ukuqaphela izwi lomuntu. Kepha izixazululo ze-biometric ezinikwe amandla ngobuhlakani bokwenziwa zingakwenza lokhu kalula. Ukuze sidizayine lesi sixazululo, sisebenzise inkundla yokubona izwi eqhathanisa izwi lofonayo ne-passive voiceprint egcinwe kusizindalwazi.
I-voiceprint yokwenziwa iwukuvela kwezibalo kwezwi lekhasimende elidalwe ngesikhathi sokusebenzelana kokuqala neplathifomu. Uma izwi lesikhathi sangempela lekhasimende lithola okufanayo ne-passive voicepass kusizindalwazi, inkundla yazisa umenzeli ukuthi ofonayo uqinisekisiwe.
Senze isixazululo sokuqaphela izwi sisebenzisa i-Phonexia (i-biometric yezwi nesofthiwe yokuqaphela inkulumo). Umsebenzi wethu ophambili ngezipikha ezintsha bekuwukwengeza amasampula ezwi azo kusizindalwazi ukuze sizikhombe futhi ekugcineni sizisebenzise ukuze siqeqeshe imodeli yokuzisiza esizwa ngezwi.
Ukuklama ukuhlonza izwi kanye nenqubo yokubhalisa kuhilele izinyathelo ezilandelayo:
2. Ukutholwa Kokukhwabanisa
Abakhwabanisi banesikhwama esigcwele amaqhinga angakhohlisa ngisho nabamele ukunakekelwa kwamakhasimende abanolwazi kakhulu. Kwabaningi babenzeli, kunzima ukubona ukuthi bayizisulu zokukhwabanisa.
Iphothali yezwi izofunda izimpawu zamazwi ezwini lofonayo, iliqhathanise nezici ezihlukile kusampuli yezwi elondolozwe ku-biometrics yezwi yokwenziwa, futhi ixwayise umenzeli mayelana nezimo ezisolisayo zocingo.
Noma yibaphi abafonayo abasolisayo bazongezwa endaweni yokugcina ehlukile. Lezi zigxivizo zezwi zizogcina zisetshenziselwe ukufundisa imodeli yokuqaphela izwi ukuthi ungayibona kanjani futhi umake ukusebenzisana okufanayo okuzayo.
Ukuzibophezela kwethu ekuqondeni izidingo zabo zephrojekthi, indlela yokubhekana nezixazululo, nokusebenzisa, okuhambisana nokusebenzelana kwethu kwangaphambili kwe-chatbot nephrojekthi yengxoxo ebukhoma, kwavula indlela yokuba iklayenti liphinde lihlanganyele nathi.
Isilinganiso sokuqala somugqa wesikhathi wephrojekthi kwakuyizinyanga ezi-3, kodwa sanwetshwa saba yizinyanga ezi-4 ngenxa yokushintshashintsha kwezidingo zamaklayenti.
Isigaba sokuthuthukisa saqala ngemihlangano yamasonto onke ukuze kwabelwane ngemininingwane ebanzi, eyabe isishintshela emihlanganweni yansuku zonke yesikram ukuze kuqinisekiswe ukuxhumana okunemininingwane phakathi kwamaqembu womabili.
Sithumele ithimba elizinikele lonjiniyela elihlanganisa -
Ithimba lethu labathuthukisi belinamakhono, lixhumana, futhi livumelana nezimo ukuze lihlangabezane nezidingo zeklayenti zephrojekthi eyimpumelelo.
Sibambe imihlangano yansuku zonke yokuqapha inqubekelaphambili futhi sasebenzisa Amaqembu e-Microsoft ukukhishwa kwesici, imibuzo yamakhasimende, nokubuyekezwa kwe-sprint, ngokuxhumana kwansuku zonke okulawulwa nge-Slack.
Sisebenzise i-Jira ukuze silandelele inqubekelaphambili esobala, sithuthukise umkhiqizo, futhi sihlele izinqubo ukuze kuhlangatshezwane nezinhloso zephrojekthi nezinto ezilethwa ngempumelelo.
Iphothali yezwi, ethuthukiswe i-Maruti Techlabs, izuzise kakhulu iklayenti lethu ngokunikeza isivikelo esisebenza ngezwi lesikhathi sangempela kanye nezinsizakalo zokuthengiselana. Lolu shintsho lube nomphumela wokwenyuka okumangalisayo okungama-36% kwezinga lokuqala lokuxazululwa kwezingcingo kanye nokuthuthuka okumangalisayo okungama-52% kumphumela wokwaneliseka kwamakhasimende.
Ukufakazela ubuqiniso obulula ngezwi lokubonwa kwebhayomethrikhi yezwi, okususa isidingo sokukhumbula amagama ayimfihlo noma imishwana ebucayi.
Ukusebenzisana kwekhasimende okulula ngokunikeza imikhonzo yomuntu siqu, ukuhlela isikhathi esimisiwe, usizo lokwenziwe, ukuxazululwa kwezikhalazo, nokuningi.
Sethule ukuqinisekiswa kobunikazi berimothi kubenzeli abasebenza bekude ukuze bavikele idatha eyimfihlo yamakhasimende nezinkampani.
Iklayenti lamukele isistimu evikelekile yezwi esekelwe ku -AI ukuze kuthuthukiswe ukuphathwa kwezingcingo zekhasimende. Lesi sixazululo sithuthukise kakhulu ukusebenza kwabo, okuyisibonelo sokuphumelela kwaso ekuphakamiseni isevisi yamakhasimende nokwaneliseka. Umphumela wensalela kube ukuqiniswa kwezinhlelo zokulwa nokukhwabanisa nokuqinisekiswa okuzenzakalelayo, okubone ukuthuthuka okukhulu kukho konke ukuzizwisa kwamakhasimende.
Silandela imikhuba ehamba phambili ye-Agile, Lean, & DevOps ukuze sakhe i-prototype ehamba phambili eyenza imibono yabasebenzisi bakho iphumelele ngokusebenzisana nokusebenzisana ngokushesha. Okuhamba phambili kwethu isikhathi sokusabela okusheshayo nokufinyeleleka.
Sifuna ngempela ukuba ithimba lakho elinwetshiwe, ngakho-ke ngaphandle kwemihlangano evamile, ungaqiniseka ukuthi ilungu ngalinye leqembu lethu linocingo olulodwa, i-imeyili, noma umlayezo.
Iklayenti belifuna ukuletha uzakwethu onobuchwepheshe nolwazi lokusebenza ngobuchwepheshe bokubona izwi. Ukuze kuzuze thina, ngaphambili sasibambisene ekwakheni nasekuletheni ngempumelelo isisombululo se-chatbot kwenye yephrojekthi yabo.
Lapho siqeda isigaba se-RFI, sifakwe ohlwini lwababambisene nabo abahlanu bokuthuthukiswa komkhiqizo abavela eGreece nase-India ku-RFP. Ngaphandle kwenqubo yokufaneleka ebekwe ngaphakathi kwe-RFI ne-RFP, i-Maruti Techlabs ifakwe ohlwini olufushane ngenxa yokuhlangenwe nakho kwayo kwangaphambilini okufanelekile ekwakheni izisombululo ezifanayo, izinkomba zamakhasimende, kanye nobudlelwane obude namakhasimende ngokwawo.
Njengoba inhloso yokuqala kuyisigaba sokutholwa enemininingwane, saziqonda ngokugcwele izidingo zekhasimende zobuchwepheshe, ukusebenza, kanye nebhizinisi. Yonke le phrojekthi yabhalwa ngokucophelela ngendlela ecacile nehlelekile, okushiya umbono omuhle kuklayenti lethu. Ngenxa yalokho, banquma ukuqhubekela phambili nesigaba sentuthuko.