Uninzi lwabantu abazama i-AI kwizimvo ze-LinkedIn® bagqibela bekwindawo enye kwezimbini. Nokuba bachitha ixesha elide behlela imveliso njengoko bebeya kuyibhala ngesandla - okanye bayivumela ukuba iposwe ngokuzenzekelayo baze bavuke bebona izimvo ezingathi azifani nabo, kwisithuba abangazange basebenze kuso.
I-AI yomntu okwi-loop (HITL) yimpendulo kuzo zombini ezi ngxaki. Idibanisa isantya se-AI kunye nokugweba komntu onedumela elisengozini. I-AI ivelisa. Uhlolo lomntu. Umntu uyayivuma. Akukho nto ithunyelwayo ngaphandle kwesigqibo esenziwe ngabom.
Ngokukodwa kwi-LinkedIn® — apho izimvo zakho ziqhotyoshelwe kwigama lakho, iprofayili yakho, kunye nophawu lwakho lobuchwephesha — le ayisiyonto imnandi ukuba nayo. Yindlela yodwa eyenza ukuphawula nge-AI kukhuseleke ukuyisebenzisa kwinqanaba eliphezulu.
I-AI ethumela imiyalezo ngaphandle kokubuza ayisosixhobo sokwenza imveliso. Luxanduva. Imodeli yomntu ngaphakathi kwendlela yokunxibelelana yiyo eyahlula ukubandakanyeka kokwenyani kwinqanaba lobukhulu kumonakalo webhrendi ngesantya.
Yintoni i-AI ye-Human-in-the-Loop kwaye kutheni ibalulekile kwi-LinkedIn®?
I-AI yomntu ngaphakathi kwi-loop ngumgaqo woyilo lwenkqubo: i-AI ivelisa imveliso, umntu uyayihlola, kwaye umntu uyayivuma ngaphambi kokuba kuthathwe naliphi na inyathelo. I-AI iphatha uhlalutyo lomthamo kunye nomxholo. Umntu ugcina igunya lokugqibela kwinto eyenzekayo ngokwenene.
Kwi-LinkedIn®, oku kubalulekile ngezizathu ezintathu ezingasebenzi kwamanye amaqonga amaninzi.
I-algorithm ye-LinkedIn® ivavanya umgangatho wezimvo ngenkuthalo. Phantsi kwe-360Brew, imodeli yokubeka umxholo ye-LinkedIn® ka-2026, izimvo ezicingisisiweyo zinobunzima obuphindwe kayi-15 kunobunzima be-algorithmic be-like - kodwa izimvo eziqhelekileyo okanye eziphindaphindwayo zihlelwa njengengxolo yokuzibandakanya kwaye zicinezelwe ngokukhutheleyo. Inkqubo ezenzekelayo ngokupheleleyo ebhala "Ukuqonda okuhle!" kwisikali ayikunyusi ukufikelela kwakho. Ibonisa kwi-algorithm ukuba ukuzibandakanya kwakho kuyasebenza - kwaye isebenza ngokufanelekileyo.
Imigaqo yeNkonzo yeLinkedIn® iyakwalela ukuthunyelwa kwemiyalezo ngokuzenzekelayo ngaphandle kokulawulwa ngumntu. I-LinkedIn® iyakwalela ngokucacileyo izixhobo ezenza izenzo ngokuzenzekelayo ngaphandle kokubandakanyeka kwamalungu. Ukuphawula ngokuzenzekelayo ngokupheleleyo - apho izithuba zesoftware egameni lakho ngaphandle kophononongo - kukwaphula loo migaqo. I-HITL ayikhuselekanga nje kuphela kwibhrendi yakho. Luyilo olukugcina uthobela imithetho.
Izimvo zakho zeLinkedIn® zinxulunyaniswa ngokusisigxina nobuwena bakho bobuchwephesha. Izimvo ezithunyelwe ngethoni engalunganga, kwisithuba esingalunganga, okanye ngempazamo yokwenyani azinyamalali buthule. Zihlala kwisithuba somnye umntu, zibonakala kwinethiwekhi yakhe, zinamathele kwigama lakho. Ingozi yodumo iphezulu apha kunakwezinye iiplatifomu.
Indlela iKonnector.ai's Human-in-the-Loop Comment Workflow esebenza ngayo:
Uphawu lwezimvo lweKonnector.ai lwakhiwe ngokupheleleyo ngokusekwe kumgaqo weHITL. Nantsi indlela echanekileyo:
Inyathelo lokuqala: I-Konnector iskena i-feed yakho ekujoliswe kuyo. Ngokusekelwe kumagama akho angundoqo acwangcisiweyo, useto lwe-ICP, kunye neeprofayili okanye amaphepha ofuna ukunxibelelana nawo, i-Konnector iveza izithuba ezifanelekileyo ezivela kubantu ofuna ukubonakala kubo. Awuphawuli ngokungacwangciswanga - uphawula ngobuchule, kwiincoko abaphulaphuli bakho ekujoliswe kubo sele benazo.
Inyathelo lesibini: I-AI ivelisa uyilo lwezimvo oluhambelana nomxholo. Ayilothemplate. Ayilotshintshiselwano lwamagama. I-AI kaKonnector ifunda umxholo weposi yokwenyani kwaye ivelise uluvo oluphendula oko kubhaliweyo - ukongeza umbono, ukubuza umbuzo olandelayo, okanye ukubhekisa kwinto ethile ethethwe ngumbhali. Ithoni kunye noseto lomntu olucwangcisiweyo kwiKonnector lubumba indlela uyilo oluvakala ngayo. Ilizwi lakho, hayi ilizwi eliqhelekileyo le-AI.
Inyathelo lesithathu: Uhlola ngaphakathi kwideshibhodi yeKonnector. Yonke idrafti ikwi-Pending queue. Uyayifunda. Ungahlela naliphi na igama, ulungise ithoni, wongeze into ethile, okanye utsibe iposi ngokupheleleyo. Akukho nto iqhubela phambili ngaphandle kwemvume yakho ecacileyo.
[Isibambi-ndawo sesikrini — Ideshibhodi yoLawulo lwezimvo zeKonnector.ai ebonisa iithebhu ezilindileyo / ezivunyiweyo / ezitsityiweyo / ezivunyiweyo ngokuzenzekelayo kunye needrafti ezintathu zezimvo ezilindele ukuphononongwa]
Inyathelo lesine: Ukuthumela izimvo ezivunyiweyo — kwaye umsebenzi uyabhalwa. Wakuba uvumile, iKonnector ithumela izimvo. Isenzo ngasinye sirekhodwa kuhlalutyo lwephulo lakho — zeziphi izithuba oye wabandakanyeka kuzo, zeziphi izimvo ezivunyiweyo, amazinga eempendulo, kunye notyelelo lweprofayili oluveliswe ngumsebenzi wakho wokuphawula. Ungabona kanye ukuba yintoni esebenzayo.
Umgca yimveliso. Ayiyo i-AI. Nabani na unokubhala izimvo. I-Konnector ikunika isiseko sokuhlaziya, ukuvuma, kunye nokulandelela kwinqanaba elithile - ngaphandle kokulahlekelwa sisigqibo esikhusela uphawu lwakho.
⚡ Yizame simahla → Seta iphulo lakho lokuqala lokuphawula ngeHITL ngemizuzu. Akukho mfuneko yekhadi letyala.
Kutheni Izimvo zeLinkedIn® eziZenzekelayo ngokupheleleyo ziyiNgcali enobungozi
Umdla wokuphawula ngokuzenzekelayo ucacile: lungiselela kube kanye, yivumele isebenze, yakha ukubonakala ngelixa ulele. Inyani ayithandeki kangako.
I-LinkedIn® iyayifumanisa. Iinkqubo zokufumanisa ze-LinkedIn® zika-2026 zihlalutya iipatheni zokuziphatha — ixesha, isantya, kunye nolwakhiwo lwezimvo — ngokwezinga. Iiakhawunti eziphawulayo kwi-milliseconds emva kokupapashwa okanye ezishiya izimvo ezifanayo ngokwesakhiwo kwizithuba ezininzi ziyavalwa. Ukufikelela obuzama ukuyakha kuyanyamalala ngokupheleleyo.
Izimvo ze-AI eziqhelekileyo ziyabonakala ngoko nangoko. “Ndiyayithanda le mbono.” “Ibaluleke kakhulu kwimeko yanamhlanje.” “Andinakuvuma okungakumbi.” La mabinzana abonakala rhoqo kwizimvo ezenziwe yi-AI kangangokuba abaphulaphuli abaziingcali be-LinkedIn® baye baphuhlisa indlela enamandla yokuqaphela iipateni zabo. Izimvo ezibonisa ukuba “i-AI ibhale oku” ayikokuzibandakanya. Yingxolo - kwaye abaphulaphuli bakho bayayazi loo nto.
Umngcipheko womonakalo webhrendi uyinyani kwaye awuthathelwa ngqalelo. Inkqubo ezenzekelayo ngokupheleleyo ayinandlela yokwazi ukuba iposi malunga nomcimbi onzima weshishini ifuna ithoni eyahlukileyo kuneyokubhiyozela inqanaba lenkampani. Ayinakutsho ukuba iposi ophawula kuyo inobuntununtunu kwezopolitiko, inempikiswano ngokomsebenzi, okanye ayifanelekanga kwisikhundla sakho. Ngaphandle kokujongwa ngumntu, umxholo owenziwe yi-AI usemngciphekweni wokuba ungabi nabuntununtunu, ungachanekanga ngokwezibakala, okanye ungabi nabuntununtunu ngokwenkcubeko - kwaye kwi-LinkedIn®, loo ngxelo ibekwe esidlangalaleni kumxholo womnye umntu.
I-HITL izisombulula zonke ezi ngxaki zintathu ngaxeshanye. Umngcipheko wokufunyanwa uyehla kuba indlela yokuziphatha ingumntu ngokwenene — uphononongo lokwenyani, utshintsho lwangempela lwexesha, izigqibo zokwenyani zokuhlela. Umgangatho wezimvo uyanda kuba umntu ufunda uyilo ngaphambi kokuba luthunyelwe. Umngcipheko wokonakala kwebhrendi uyasuswa kuba akukho nto iphilayo ngaphandle kotyikityo lwakho.
Yintoni Eyenza Ingxelo yeLinkedIn® Esemgangathweni Ophezulu - kunye nendlela i-AI kaKonnector eyifanisa ngayo
Phantsi kwe-360Brew, imisonto yezimvo eneempendulo ezintathu nangaphezulu ivelisa isiphumo sokukhulisa i-5.2x ekusasazweni komxholo. Oko kwenzeka kuphela xa izimvo zokuqala zilungile ngokwaneleyo ukuba zibangele impendulo evela kumbhali weposi okanye kwabanye abafundi.
I-AI kaKonnector yenzelwe ukuvelisa izimvo ezihlangabezana nalo mgangatho — kungekhona izimvo ezikhoyo kuphela.
Ingxelo ye-LinkedIn® esemgangathweni ophezulu yenza enye yezinto ezintlanu: yongeza umbono ongafakwanga kwisithuba, ibuza umbuzo olandelayo, ibhekisa kwinto ethile ecacileyo kumxholo wesithuba, yabelana ngamava obuqu afanelekileyo ngokuthe ngqo, okanye inika inqaku ledatha elandisa ingxoxo yombhali.
Oko ingakwenziyo: vumelana ngokubanzi, shwankathela isithuba sombhali, okanye sebenzisa uhlobo lwamabinzana agcwalisayo i-360Brew ewafundileyo ukuwahlulahlula njengengxolo.
Useto lwethoni lukaKonnector lukuvumela ukuba ulungiselele indlela i-AI evakala ngayo — ehlalutyayo ngakumbi, ethetha ngakumbi, ephezulu ngakumbi, ethe ngqo ngakumbi — ukuze iidrafti ezivelisayo zivakale njengenguqulelo yakho, kungekhona inguqulelo yomzekelo omkhulu wolwimi wokubhala wonke umntu ngaxeshanye.
Ubude obufanelekileyo bubalulekile. Ayilocala elinye. Ayilobali. Izivakalisi ezibini ukuya kwezine ezongeza into ethile - yifomathi efumana iimpendulo, ivelise utyelelo lweprofayili, kwaye yakha uhlobo lokubonakala oluhambelanayo ngokuhamba kwexesha.
Ukuphawula okuSebenzayo kwi-Human-in-the-Loop vs. Okuzenzakalelayo ngokupheleleyo
| Criteria | Umntu okwiLuphu (Konnector.ai) | Zizenzekelayo ngokuzenzekelayo |
|---|---|---|
| Ukhuseleko lwebhrendi | Phezulu — uyayivuma yonke ingxelo ngaphambi kokuba iposwe | Iphantsi — akukho phononongo ngaphambi kokuba iposwe; umngcipheko ongekho lizwi uhlala uhleli |
| Ukuthobela i-LinkedIn® ToS | Ukuthobela — ukujongwa kwabantu kuyaqhubeka kulo lonke ixesha | Usemngciphekweni — ukuthunyelwa ngokuzenzekelayo kunokubangela iiflegi zomgaqo-nkqubo |
| Umgangatho wezimvo | Imixholo kunye neyomntu ngamnye — iidrafti ze-AI, ukuphuculwa kwabantu | Idla ngokuba yinto eqhelekileyo okanye ephindaphindwayo — ibonwa yi-algorithm kunye nabaphulaphuli |
| Ixesha ligciniwe | Okubalulekileyo — I-AI ibhala izimvo ezingama-50 ngexesha oza kubhala ngalo ezi-5 | Eyona nto iphezulu — kodwa ngexabiso lolawulo nomgangatho |
| Umngcipheko wokubhaqwa | Low — uphononongo lwabantu ludala umahluko wokwenyani wexesha | Phezulu - iiakhawunti ziphawulwe ngesantya esingaphezu kwamandla omntu kunye nokuphindaphinda kwepateni |
| ulawulo | Igcwele — hlela okanye ulahle naluphi na uyilo nangaliphi na ixesha | Akukho nto — xa ulandelelwano luqhubeka, luyaqhuba |
| Impembelelo yokufikelela | Ezilungileyo — izimvo ezisemgangathweni zifumana iimpendulo kwaye ziqhubele phambili ukutyelela iiprofayili | Ukungathathi cala ukuya kokungalunganga — izimvo eziqhelekileyo zihlelwe njengengxolo yokuzibandakanya phantsi kwe-360Brew |
Iingcali ezivelisa izimvo ezininzi zeLinkedIn® ngo-2026 azizizo ezithumela kakhulu. Zizo ezithumela izimvo zazo zilungile ngokwaneleyo ukwenza abantu bafune ukwazi ukuba ngubani owazibhalayo.
Ukuze ufumane okungakumbi malunga nendlela ukuphawula nge-AI okuhambelana ngayo nesicwangciso sokuzibandakanya se-LinkedIn® esithobela imithetho, jonga izikhokelo zethu malunga nezimvo ze-CFBR kunye ne-AI kwi-LinkedIn® kunye nendlela izimvo ze-AI ezikhuthaza ngayo ukubandakanyeka kwe-LinkedIn®.
📅 Bhukisha iDemo yasimahla → Jonga indlela yokusebenza kwezimvo ze-HITL yeKonnector.ai ngqo — ukusuka kwisithuba seposi ukuya kuphononongo oluyilwayo ukuya kukubandakanyeka okuvunyiweyo.
⚡ Bhalisa simahla → Hlola ngaphambi kokuba uthumele. Hlala uyinyani kwinqanaba. Qala namhlanje.
Ukufikelela kwakho kwi-LinkedIn® kube yi-11x
Ukuzenzekela kunye neGen AI
Sebenzisa amandla e-LinkedIn® Automation kunye ne-Gen AI ukuze wandise ukufikelela kwakho ngendlela ongazange uyifikelele ngaphambili. Bandakanya amawaka abantu abakhokelayo veki nganye ngezimvo eziqhutywa yi-AI kunye neephulo ezijoliswe kuzo—zonke zivela kwiqonga elinye elinamandla le-lead-gen.
Imibuzo ebuzwa qho
I-AI ye-Human-in-the-loop (HITL) yeengxelo ze-LinkedIn® yinkqubo apho i-AI ivelisa iidrafti zezimvo, kodwa umntu uyazihlola aze azivume ngaphambi kokuba azithumele. Oku kuqinisekisa ukuchaneka, ulawulo lwethoni, kunye nokhuseleko lwebhrendi ngelixa kusagcinwa ixesha.
Ewe—ukuba usebenzisa indlela yomntu ngaphakathi kwe-loop. Ukuphawula nge-AI ngokuzenzekelayo ngokupheleleyo kunokukhokelela kwiimpendulo eziqhelekileyo, ukwaphulwa komgaqo-nkqubo, okanye umngcipheko wodumo. I-HITL iqinisekisa ukuba yonke ingxelo iphononongwa ngaphambi kokuba ivele.
I-LinkedIn® ivumela umsebenzi oncediswa yi-AI kuphela xa kukho ukubandakanyeka komntu. Ukuthumela ngokuzenzekelayo ngaphandle kophononongo lomsebenzisi kunokwaphula iMigqaliselo yeNkonzo ye-LinkedIn® kwaye kunokubangela imiqathango.
Izimvo ezizenzekelayo zihlala zingenawo umxholo, ziphinda-phinda iipateni, kwaye zithumela ngesantya esingaqhelekanga. Oku kwenza kube lula kwiinkqubo zeLinkedIn® ukuba zibhaqwe kwaye kunokukhokelela ekunciphiseni ukufikelela, ukuvalwa kwesithunzi, okanye imiqathango yeakhawunti.
I-HITL AI iphucula ukubandakanyeka ngokudibanisa isantya se-AI kunye nombono womntu. Oku kuphumela kwizimvo ezifanelekileyo nezicingisisiweyo ezinokuthi zifumane iimpendulo kunye nokutyelela iiprofayili.
Ingxelo esemgangathweni ophezulu yongeza ixabiso. Ingabandakanya umbono owahlukileyo, ukubuza umbuzo ofanelekileyo, ukubhekisa kwiingongoma ezithile kwisithuba, okanye ukwabelana ngamava anxulumene noko—kunokusebenzisa amabinzana aqhelekileyo.
Ewe, ukuba zisetyenziswe ngaphandle kophononongo. Izimvo ezibhalwe kakubi okanye ezingabalulekanga zinokubonakala zingachanekanga kwaye zonakalise ukuthembeka. Ekubeni izimvo zeLinkedIn® zisesidlangalaleni kwaye zihlala zihleli, zonke izimvo zibonisa ubuwena bakho bobuchwephesha.
Umgangatho ubaluleke ngaphezu kobuninzi. Ukuthumela izimvo ezi-5–10 ezicingisisiweyo nezifanelekileyo imihla ngemihla kusebenza ngakumbi kunokuthumela iimpendulo ezininzi ezikumgangatho ophantsi okanye ezizenzekelayo.
Ukuphawula okuncediswa yi-AI kubandakanya ukuvelisa iidrafti eziphononongwa kwaye zivunywe ngumntu. Izithuba zokuphawula ezizenzekelayo ngokuthe ngqo ngaphandle kwegalelo lomntu. Eyokuqala ikhuselekile kwaye iyavumelana; eyesibini inomngcipheko.
Eyona ndlela ikhuselekileyo yokwandisa ulwazi kukusebenzisa i-AI ukuvelisa iidrafti kunye nomntu oza kuzihlola ngaphambi kokuba azithumele. Oku kugcina ubunyani, kuqinisekisa ukuthotyelwa kwemithetho, kwaye kukhusela udumo lwakho lobungcali.
