中(zhong)國(guo)人民銀行(xing)《金(jin)融科技發展規(gui)劃(hua)（2022~2025年》指(zhi)出(chu)：運用機器人流(liu)程(cheng)自(zi)動化(hua)（RPA）、自(zi)然語言處理(li)（NLP）、智能字符識別（ICR）等智能技術開展(zhan)端到端數(shu)字化(hua)流(liu)程(cheng)重構，打通部門間業(ye)(ye)務阻隔(ge)與流(liu)程(cheng)斷(duan)點，實(shi)現跨角色、跨時(shi)序的業(ye)(ye)務靈活(huo)定制與編排，打造環節(jie)無縫銜接、信息實(shi)時(shi)交(jiao)互(hu)、資源協同高效的業(ye)(ye)務處理(li)模式，更好支(zhi)撐數(shu)字化(hua)業(ye)(ye)務發展(zhan)。

人(ren)保財險(xian)科技(ji)(ji)運營部(bu)(bu)自2019年(nian)(nian)以(yi)來探索引入RPA技(ji)(ji)術，打(da)(da)造數(shu)字員(yuan)工(gong)(gong)(gong)(gong)，將大(da)量重復且規(gui)則明(ming)確的(de)業(ye)務(wu)操(cao)作(zuo)由機器(qi)人(ren)自動執行替代人(ren)工(gong)(gong)(gong)(gong)操(cao)作(zuo)，節省員(yuan)工(gong)(gong)(gong)(gong)在機械、重復、耗(hao)時(shi)工(gong)(gong)(gong)(gong)作(zuo)上(shang)的(de)工(gong)(gong)(gong)(gong)作(zuo)量和(he)(he)時(shi)間，使員(yuan)工(gong)(gong)(gong)(gong)可以(yi)從事更有價(jia)值(zhi)和(he)(he)創(chuang)造性的(de)工(gong)(gong)(gong)(gong)作(zuo)，有效提(ti)升運營效能和(he)(he)人(ren)力績效。經過兩年(nian)(nian)左右的(de)不斷嘗(chang)試并積(ji)累(lei)經驗，2022年(nian)(nian)，科技(ji)(ji)運營部(bu)(bu)將數(shu)字員(yuan)工(gong)(gong)(gong)(gong)平(ping)臺項目納(na)入人(ren)保財險(xian)“三化”建(jian)設首批項目清(qing)單，通過建(jian)章立制、組建(jian)CoE（Center of Excellence，一般稱卓越中心，負責統(tong)籌推(tui)動RPA應用）、深挖(wa)場景、打(da)(da)造平(ping)臺、強化管控、技(ji)(ji)術創(chuang)新等舉措，自上(shang)而下有組織(zhi)地推(tui)動RPA系統(tong)化、規(gui)范(fan)化、規(gui)模化應用，促(cu)進降本增(zeng)效，助力公司運營自動化和(he)(he)數(shu)字化轉(zhuan)型(xing)。

截至2022年(nian)11月底(di)，數(shu)字(zi)員(yuan)工(gong)(gong)平臺已在人(ren)(ren)保財(cai)險總分(fen)公司上(shang)線612個RPA流程，打造699名數(shu)字(zi)員(yuan)工(gong)(gong)，涵蓋總部及33家(jia)省(sheng)級分(fen)公司。2022年(nian)以來，數(shu)字(zi)員(yuan)工(gong)(gong)累計(ji)運(yun)行時長超過84萬小時，節約人(ren)(ren)工(gong)(gong)工(gong)(gong)作量超過356人(ren)(ren)年(nian)，有效(xiao)提升了業務運(yun)營效(xiao)率和人(ren)(ren)力(li)效(xiao)能，實現降本增效(xiao)。

廣泛(fan)應(ying)用的(de)車險核(he)保(bao)機(ji)器人。在車(che)險(xian)業務中(zhong)，存在大量(liang)規則清(qing)晰且重復性高的(de)核(he)保(bao)(bao)工(gong)(gong)(gong)作(zuo)(zuo)，核(he)保(bao)(bao)處理(li)(li)(li)時需(xu)要(yao)操作(zuo)(zuo)多個系統(tong)，人(ren)(ren)(ren)(ren)(ren)工(gong)(gong)(gong)處理(li)(li)(li)的(de)工(gong)(gong)(gong)作(zuo)(zuo)量(liang)、復雜度(du)、易錯程度(du)都較高，處理(li)(li)(li)的(de)時效性也不能完全滿足(zu)業務需(xu)要(yao)。針對這一(yi)場景，開發車(che)險(xian)核(he)保(bao)(bao)機器人(ren)(ren)(ren)(ren)(ren)輔助(zhu)人(ren)(ren)(ren)(ren)(ren)工(gong)(gong)(gong)處理(li)(li)(li)核(he)保(bao)(bao)工(gong)(gong)(gong)作(zuo)(zuo)，通過RPA融合(he)OCR技術(shu)，提升信息審核(he)速度(du)及(ji)準確性，輔助(zhu)人(ren)(ren)(ren)(ren)(ren)工(gong)(gong)(gong)快速完成(cheng)核(he)保(bao)(bao)操作(zuo)(zuo)，效率(lv)提升600%。采用總分(fen)協同(tong)模式(shi)對數字員工(gong)(gong)(gong)進行集中(zhong)管理(li)(li)(li)，目前車(che)險(xian)核(he)保(bao)(bao)機器人(ren)(ren)(ren)(ren)(ren)已(yi)在28家省級(ji)分(fen)公(gong)司上線，2022年以(yi)來已(yi)累計節約37人(ren)(ren)(ren)(ren)(ren)年工(gong)(gong)(gong)作(zuo)(zuo)量(liang)。其中(zhong)，人(ren)(ren)(ren)(ren)(ren)保(bao)(bao)財險(xian)福建(jian)分(fen)公(gong)司共上線9個車(che)險(xian)核(he)保(bao)(bao)機器人(ren)(ren)(ren)(ren)(ren)，平均每天處理(li)(li)(li)300單業務，替代7名人(ren)(ren)(ren)(ren)(ren)工(gong)(gong)(gong)。

集中部署(shu)的理賠集審照(zhao)片分(fen)揀機器(qi)人。在理(li)(li)賠(pei)業(ye)務中(zhong)，在線(xian)理(li)(li)賠(pei)坐(zuo)席需要(yao)通(tong)過(guo)集(ji)中(zhong)審核(he)平臺對(dui)各個渠道(dao)上(shang)傳的(de)海量照(zhao)(zhao)片(pian)進行(xing)(xing)人(ren)(ren)(ren)工(gong)審核(he)和(he)分(fen)揀，工(gong)作量大，耗費人(ren)(ren)(ren)力(li)成(cheng)本高。針對(dui)這一場景，開發理(li)(li)賠(pei)集(ji)審照(zhao)(zhao)片(pian)分(fen)揀機器(qi)人(ren)(ren)(ren)，通(tong)過(guo)RPA融合OCR技術(shu)對(dui)照(zhao)(zhao)片(pian)關鍵(jian)信(xin)息進行(xing)(xing)自動提取和(he)識(shi)別，將(jiang)識(shi)別后的(de)照(zhao)(zhao)片(pian)自動分(fen)揀至對(dui)應目錄，大幅(fu)提升線(xian)上(shang)理(li)(li)賠(pei)處理(li)(li)時效(xiao)(xiao)(xiao)。理(li)(li)賠(pei)集(ji)審照(zhao)(zhao)片(pian)分(fen)揀機器(qi)人(ren)(ren)(ren)自2021年年底(di)上(shang)線(xian)以來，采用總(zong)部集(ji)中(zhong)部署的(de)方式，共(gong)部署15名數字(zi)員(yuan)工(gong)，累計運行(xing)(xing)22019小時，節(jie)約人(ren)(ren)(ren)力(li)1633人(ren)(ren)(ren)天，有(you)效(xiao)(xiao)(xiao)提高了分(fen)揀效(xiao)(xiao)(xiao)率和(he)準確率，在為(wei)理(li)(li)賠(pei)坐(zuo)席減(jian)負的(de)同時，將(jiang)數字(zi)員(yuan)工(gong)的(de)處理(li)(li)結(jie)果記錄在表格(ge)中(zhong)，每(mei)日交由(you)指定人(ren)(ren)(ren)員(yuan)對(dui)結(jie)果進行(xing)(xing)二次檢查，保證流程運行(xing)(xing)平穩且風險可(ke)控(kong)。

處理短期大批量業(ye)務的農險業(ye)務機(ji)器人。農(nong)險(xian)(xian)(xian)業(ye)(ye)(ye)務(wu)具有受(shou)政(zheng)策約(yue)束較大(da)(da)、集(ji)中出單占比高、季節性(xing)較強等特(te)點。在一些特(te)殊的應用場景，如財政(zheng)部下發(fa)《中央財政(zheng)農(nong)業(ye)(ye)(ye)保(bao)險(xian)(xian)(xian)保(bao)費(fei)補貼(tie)(tie)管理(li)辦(ban)法》要求(qiu)調整政(zheng)策性(xing)農(nong)險(xian)(xian)(xian)各(ge)險(xian)(xian)(xian)種的各(ge)級財政(zheng)補貼(tie)(tie)比例(li)、農(nong)險(xian)(xian)(xian)集(ji)中承保(bao)錄單時進行海量(liang)(liang)影像上(shang)傳、農(nong)險(xian)(xian)(xian)理(li)賠大(da)(da)客戶案件(jian)(jian)抽(chou)檢要求(qiu)下載(zai)系統案件(jian)(jian)清單及(ji)影像資(zi)料(liao)(liao)時，業(ye)(ye)(ye)務(wu)量(liang)(liang)巨(ju)大(da)(da)且(qie)對時效(xiao)性(xing)及(ji)數(shu)據(ju)(ju)質量(liang)(liang)要求(qiu)都很高，分公(gong)司(si)往(wang)往(wang)需要抽(chou)調大(da)(da)量(liang)(liang)業(ye)(ye)(ye)務(wu)人(ren)(ren)員集(ji)中在數(shu)天或數(shu)十天內加班(ban)加點完成上(shang)萬件(jian)(jian)保(bao)單的錄單、批(pi)改(gai)或資(zi)料(liao)(liao)下載(zai)操(cao)作。為(wei)提高業(ye)(ye)(ye)務(wu)處(chu)理(li)效(xiao)率，開發(fa)上(shang)線農(nong)險(xian)(xian)(xian)政(zheng)策性(xing)系數(shu)批(pi)改(gai)機(ji)(ji)器(qi)人(ren)(ren)、農(nong)險(xian)(xian)(xian)影像上(shang)傳機(ji)(ji)器(qi)人(ren)(ren)、農(nong)險(xian)(xian)(xian)理(li)賠資(zi)料(liao)(liao)下載(zai)機(ji)(ji)器(qi)人(ren)(ren)等農(nong)險(xian)(xian)(xian)業(ye)(ye)(ye)務(wu)機(ji)(ji)器(qi)人(ren)(ren)。根據(ju)(ju)業(ye)(ye)(ye)務(wu)需要，快速(su)部署(shu)(shu)數(shu)十名數(shu)字(zi)(zi)員工(gong)，完成短期大(da)(da)批(pi)量(liang)(liang)任(ren)務(wu)后再將機(ji)(ji)器(qi)人(ren)(ren)許可資(zi)源釋放，極大(da)(da)地發(fa)揮了(le)數(shu)字(zi)(zi)員工(gong)快速(su)部署(shu)(shu)、靈活適配業(ye)(ye)(ye)務(wu)的優勢，減(jian)輕(qing)了(le)基(ji)層(ceng)一線的工(gong)作負擔。2022年7月(yue)(yue)，人(ren)(ren)保(bao)財險(xian)(xian)(xian)河(he)南分公(gong)司(si)緊急開發(fa)上(shang)線并部署(shu)(shu)15個農(nong)險(xian)(xian)(xian)政(zheng)策性(xing)系數(shu)批(pi)改(gai)機(ji)(ji)器(qi)人(ren)(ren)，共(gong)運(yun)行2個月(yue)(yue)，完成23890單批(pi)改(gai)工(gong)作，節約(yue)人(ren)(ren)力498人(ren)(ren)天。

開創數字(zi)員工(gong)服務化模式的運維機器人。人(ren)保財險湖(hu)南(nan)(nan)分公(gong)司科(ke)技(ji)運營(ying)部率先提出“數(shu)(shu)字(zi)(zi)員(yuan)(yuan)工(gong)(gong)(gong)服務化(hua)”理念并(bing)實(shi)現(xian)落地，自(zi)(zi)主開發“小(xiao)(xiao)南(nan)(nan)同(tong)學”智(zhi)能(neng)運維(wei)應用。一(yi)方面(mian)依托企(qi)業微信工(gong)(gong)(gong)作(zuo)(zuo)臺作(zuo)(zuo)為(wei)入(ru)(ru)口(kou)，接(jie)收用戶以(yi)自(zi)(zi)然語言方式下(xia)達的任務命(ming)令并(bing)轉化(hua)為(wei)數(shu)(shu)字(zi)(zi)員(yuan)(yuan)工(gong)(gong)(gong)可以(yi)理解的任務消(xiao)息(xi)，另(ling)一(yi)方面(mian)將各(ge)類(lei)運維(wei)數(shu)(shu)字(zi)(zi)員(yuan)(yuan)工(gong)(gong)(gong)封裝成標準化(hua)服務以(yi)便調用。各(ge)數(shu)(shu)字(zi)(zi)員(yuan)(yuan)工(gong)(gong)(gong)監聽消(xiao)息(xi)，當監聽到(dao)歸屬于(yu)自(zi)(zi)身的任務消(xiao)息(xi)時則自(zi)(zi)動執行(xing)，并(bing)將執行(xing)結果(guo)返回(hui)“小(xiao)(xiao)南(nan)(nan)同(tong)學”前端并(bing)反(fan)饋(kui)用戶。“小(xiao)(xiao)南(nan)(nan)同(tong)學”將日常(chang)運維(wei)中的重復工(gong)(gong)(gong)作(zuo)(zuo)交給數(shu)(shu)字(zi)(zi)員(yuan)(yuan)工(gong)(gong)(gong)自(zi)(zi)動解決，如車險投保預確(que)認、重新轉保單、電(dian)子保單重發等，實(shi)現(xian)運維(wei)工(gong)(gong)(gong)作(zuo)(zuo)的智(zhi)能(neng)化(hua)自(zi)(zi)執行(xing)，并(bing)將數(shu)(shu)字(zi)(zi)員(yuan)(yuan)工(gong)(gong)(gong)的調用無縫融入(ru)(ru)運維(wei)工(gong)(gong)(gong)作(zuo)(zuo)場景，支(zhi)持數(shu)(shu)字(zi)(zi)員(yuan)(yuan)工(gong)(gong)(gong)的自(zi)(zi)定義擴展和快速接(jie)入(ru)(ru)，實(shi)現(xian)“數(shu)(shu)字(zi)(zi)員(yuan)(yuan)工(gong)(gong)(gong)服務化(hua)”模式。人(ren)保財險科(ke)技(ji)運營(ying)部基于(yu)湖(hu)南(nan)(nan)分公(gong)司創新探索，打造(zao)統頒(ban)“人(ren)保小(xiao)(xiao)智(zhi)”智(zhi)能(neng)運維(wei)應用并(bing)進(jin)行(xing)全(quan)國推廣(guang)。

數字(zi)員(yuan)(yuan)(yuan)(yuan)工(gong)(gong)(gong)具有廣(guang)闊的應(ying)(ying)用(yong)前(qian)景。下一步，人保(bao)財險(xian)科技運營部將(jiang)持續深化數字(zi)員(yuan)(yuan)(yuan)(yuan)工(gong)(gong)(gong)建設(she)和(he)應(ying)(ying)用(yong)，協同業務(wu)部門及分公司進一步挖掘、擴大應(ying)(ying)用(yong)場景，優化、完善(shan)數字(zi)員(yuan)(yuan)(yuan)(yuan)工(gong)(gong)(gong)運營管理平(ping)臺支(zhi)撐能力(li)，提升數字(zi)員(yuan)(yuan)(yuan)(yuan)工(gong)(gong)(gong)效(xiao)能以及運行(xing)穩定性，加強總分協同、技術培(pei)訓和(he)流程(cheng)推廣(guang)，深化RPA與AI等技術融合創新，積(ji)極(ji)與同業交流學(xue)習先進經驗，探索應(ying)(ying)用(yong)RPM（機器(qi)人流程(cheng)挖掘）技術發掘更多應(ying)(ying)用(yong)場景，進一步促進降本增效(xiao)，助力(li)運營向(xiang)自動(dong)化和(he)智能化轉型。