Co esponding au ho : Su esh Kuma Maddala
Copy igh © 2025 Au ho (s) e ain he copy igh o his a icle. This a icle is published unde he e ms o he C ea i e Commons A ibu ion Liscense 4.0.
AI-d i en pe sonaliza ion in consume goods and e ail: A echnical analysis
Su esh Kuma Maddala *
Uni e si y o Hyde abad, India.
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 458-463
Publica ion his o y: Recei ed on 25 Ma ch 2025; e ised on 02 May 2025; accep ed on 04 May 2025
A icle DOI: h ps://doi.o g/10.30574/wja .2025.26.2.1639
Abs ac
AI-d i en pe sonaliza ion has become a c i ical compe i i e ad an age in mode n e ail en i onmen s, enabling ailo ed
cus ome expe iences ac oss digi al and physical ouchpoin s. This a icle explo es he ans o ma i e ole o a i icial
in elligence echnologies in eshaping consume goods and e ail pe sonaliza ion s a egies. Beginning wi h an
o e iew o undamen al AI pe sonaliza ion echnologies, he discussion p og esses h ough ad anced
ecommenda ion engine a chi ec u es, dynamic p icing implemen a ions, con e sa ional AI sys ems, and in-s o e
pe sonaliza ion solu ions. The a icle examines how hese echnologies c ea e cohesi e pe sonalized expe iences ha
inc ease engagemen , d i e sales, and os e cus ome loyal y while add essing echnical challenges and
implemen a ion conside a ions o e aile s na iga ing he e ol ing digi al comme ce landscape.
Keywo ds: Pe sonaliza ion; Recommenda ion Sys ems; Dynamic P icing; Con e sa ional AI; Augmen ed Reali y
1. In oduc ion
In he mode n e ail landscape, pe sonaliza ion has eme ged as a c i ical compe i i e di e en ia o . Today's consume s
expec ailo ed ecommenda ions, cus omized shopping expe iences, and seamless in e ac ions ac oss bo h online and
o line channels. A i icial In elligence (AI) is ans o ming pe sonaliza ion capabili ies by le e aging machine lea ning,
deep lea ning, and ad anced da a analy ics o deli e eal- ime, highly ele an cus ome expe iences ac oss he e ail
alue chain.
Acco ding o McKinsey, e aile s using pe sonaliza ion s a egies ha e seen e enue inc eases o 15-20% and cos
educ ions o 10-30% compa ed o compe i o s who don' implemen hese echnologies [1]. This subs an ial impac
s ems om AI's abili y o p ocess as amoun s o consume da a o sugges ele an p oduc s a p ecisely he igh
momen . Majo pla o ms now analyze e aby es o cus ome in e ac ion da a daily, c ea ing sophis ica ed p e e ence
models ha adap in eal- ime o changing consume beha io s.
Recen esea ch in he ield o human-compu e in e ac ion has e ealed ha consume s exhibi signi ican ly highe
engagemen le els wi h AI-pe sonalized in e aces. In a comp ehensi e s udy examining use esponses o pe sonalized
e ail expe iences, esea che s ound ha cus omized ecommenda ions inc eased pu chase in en by 35% and
imp o ed o e all sa is ac ion sco es by 28% compa ed o gene ic in e aces [2]. This e ec was pa icula ly p onounced
when pe sonaliza ion ex ended beyond p oduc ecommenda ions o include cus omized na iga ion pa hways and
indi idualized p omo ional con en .
The in eg a ion o AI ac oss he e ail alue chain has e olu ionized how businesses unde s and and espond o
cus ome needs. By syn hesizing da a om mul iple ouchpoin s—including b owsing his o y, pu chase pa e ns, and
in-s o e beha io —AI c ea es uni ied cus ome p o iles ha enable seamless expe iences ac oss channels. This
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 458-463
459
capabili y has become essen ial in an e a whe e 73% o shoppe s use mul iple channels du ing hei pu chasing jou ney
[1].
2. Co e AI Technologies Powe ing Re ail Pe sonaliza ion
The ounda ion o AI-d i en pe sonaliza ion consis s o se e al key echnologies ha collec i ely ans o m how
e aile s unde s and and engage wi h cus ome s. These in e connec ed sys ems le e age massi e da a se s o c ea e
inc easingly p ecise cus ome expe iences ac oss mul iple ouchpoin s.
Cus ome segmen a ion algo i hms ha e e ol ed signi ican ly beyond adi ional demog aphic-based app oaches.
Mode n AI-d i en segmen a ion can now iden i y complex beha io al pa e ns by analyzing housands o cus ome
in e ac ions simul aneously. Acco ding o ma ke esea ch, e aile s implemen ing ad anced AI segmen a ion
echniques ha e achie ed up o 30% imp o emen in ma ke ing ROI compa ed o adi ional segmen a ion me hods
[3]. These sys ems excel a iden i ying no jus who cus ome s a e, bu p edic ing wha hey' e likely o do nex . Fo
ins ance, AI segmen a ion can now p edic wi h ema kable accu acy which cus ome s a e abou o chu n o which
segmen s a e mos ecep i e o speci ic p oduc ca ego ies, allowing e aile s o de elop a ge ed e en ion s a egies
ha ha e shown o educe cus ome a i ion by 20-25% in compe i i e ma ke s.
Recommenda ion sys ems ha e become inc easingly sophis ica ed, inco po a ing mul iple modeling app oaches o
deli e highly ele an sugges ions. Resea ch published in Symme y demons a es ha hyb id ecommenda ion
engines— hose combining collabo a i e il e ing wi h con en -based app oaches—consis en ly ou pe o m single-
me hod sys ems, achie ing up o 27% highe accu acy in p edic ing cus ome p e e ences [4]. These sys ems analyze
pa e ns ac oss simila use s while simul aneously e alua ing indi idual beha io pa e ns o c ea e highly
pe sonalized ecommenda ions. Amazon's ecommenda ion engine exempli ies his app oach, le e aging deep lea ning
echniques o analyze billions o in e ac ion da a poin s daily and con ibu ing an es ima ed 35% o he company's o al
sales.
Dynamic p icing engines now inco po a e eal- ime compe i i e analysis wi h demand o ecas ing o op imize p ice
poin s. These sys ems con inuously moni o ma ke condi ions, in en o y le els, and compe i o p icing o adjus o e s
dynamically. Machine lea ning algo i hms can p edic op imal p ice poin s ac oss housands o SKUs simul aneously,
wi h e aile s epo ing a e age ma gin imp o emen s be ween 3-5% a e implemen a ion [3].
Pe sonalized ma ke ing op imiza ion pla o ms u ilize ad anced p edic i e analy ics o de e mine which messages will
esona e wi h speci ic cus ome s. As no ed in esea ch published in Symme y, AI-powe ed ma ke ing sys ems can
inc ease engagemen a es by up o 40% compa ed o non-pe sonalized app oaches by deli e ing he igh con en
h ough he igh channel a he igh ime [4]. These sys ems lea n om cus ome esponse pa e ns, con inuously
op imizing message iming, channel selec ion, and con en o maximize engagemen p obabili y.
Table 1 AI Re ail Pe sonaliza ion Pe o mance Me ics [3, 4]
AI Technology
Key Pe o mance Indica o
Imp o emen Pe cen age
Cus ome Segmen a ion Algo i hms
Ma ke ing ROI
30%
Cus ome Segmen a ion Algo i hms
Cus ome A i ion Reduc ion
20-25%
Recommenda ion Sys ems
P edic ion Accu acy
27%
Recommenda ion Sys ems
Con ibu ion o To al Sales (Amazon)
35%
Dynamic P icing Engines
Ma gin Imp o emen
3-5%
Pe sonalized Ma ke ing Pla o ms
Engagemen Ra e Inc ease
40%
3. Ad anced Recommenda ion Engine A chi ec u es
Recommenda ion engines se e as he co ne s one o AI-d i en pe sonaliza ion in e ail, wi h h ee p ima y
a chi ec u al app oaches ha ha e demons a ed signi ican impac on cus ome engagemen and e enue gene a ion.
These sophis ica ed sys ems p ocess as quan i ies o da a o deli e inc easingly p ecise ecommenda ions ha d i e
pu chasing decisions.
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 458-463
460
Collabo a i e il e ing sys ems analyze pa e ns ac oss use beha io s o iden i y simila i ies be ween cus ome s and
make ecommenda ions based on hese ela ionships. Acco ding o esea ch published in Applied Sciences, hese
sys ems can be ca ego ized in o memo y-based and model-based app oaches, wi h ma ix ac o iza ion echniques
showing pa icula p omise in la ge-scale deploymen s. A comp ehensi e e alua ion ac oss mul iple domains e ealed
ha collabo a i e il e ing implemen a ions can achie e ecommenda ion p ecision a es o up o 87.5% when p ope ly
op imized [5]. These sys ems excel in en i onmen s wi h ich use in e ac ion da a, making hem pa icula ly e ec i e
o es ablished e-comme ce pla o ms. Howe e , hey s uggle wi h he "cold s a " p oblem when aced wi h new use s
o i ems—a signi ican limi a ion ha has d i en esea ch in o al e na i e app oaches.
Con en -based il e ing sys ems analyze he a ibu es o p oduc s and ma ch hem wi h use p e e ences, c ea ing
di ec ela ionships be ween i em cha ac e is ics and use a ini y. These sys ems excel in con ex s whe e i em
a ibu es a e well-de ined and ex ensi e. Ne lix's implemen a ion o con en -based il e ing has become he indus y
s anda d, ca ego izing con en along housands o a ibu es o c ea e ema kably p ecise ma ching be ween iewing
his o y and new con en ecommenda ions. Resea ch indica es ha ad anced con en -based sys ems can achie e up o
76% accu acy in p edic ing use p e e ences based solely on i em a ibu es [5]. The p ima y ad an age o hese
sys ems is hei independence om o he use s' da a, making hem highly e ec i e o niche p oduc s and specialized
ecommenda ions.
Hyb id ecommenda ion sys ems combine mul iple app oaches o o e come he limi a ions inhe en in single-me hod
a chi ec u es. As no ed in comp ehensi e s udies o deep lea ning-based ecommenda ion sys ems, hyb id models
consis en ly ou pe o m single-app oach me hods by le e aging complemen a y s eng hs [6]. These sys ems can
signi ican ly educe he cold-s a p oblem while main aining high ecommenda ion quali y o es ablished use s.
Spo i y exempli ies his app oach, employing a sophis ica ed ensemble o models ha inco po a e bo h collabo a i e
pa e ns and con en analysis. Resea ch indica es ha hyb id sys ems can imp o e ecommenda ion quali y by 15-20%
compa ed o single-app oach me hods ac oss mul iple e alua ion me ics [6]. Mode n hyb id a chi ec u es inc easingly
inco po a e deep lea ning echniques, wi h neu al collabo a i e il e ing and a en ion mechanisms showing pa icula
p omise in cap u ing complex ela ionships be ween use s and i ems.
Table 2 Pe o mance Me ics o Re ail Recommenda ion Engine A chi ec u es [5, 6]
Recommenda ion
Sys em A chi ec u e
Accu acy/Pe o mance
Me ic
Pe cen age
Key S eng h
No able
Implemen a ion
Collabo a i e Fil e ing
Sys ems
Recommenda ion
P ecision Ra e
87.5%
Excels wi h ich
use in e ac ion
da a
La ge-scale e-
comme ce pla o ms
Con en -Based Fil e ing
Sys ems
P edic ion Accu acy
76%
Independence om
o he use s' da a
Ne lix con en
ecommenda ions
Hyb id
Recommenda ion
Sys ems
Imp o emen O e Single
Me hods
15-20%
Reduces cold-s a
p oblem
Spo i y pe sonalized
playlis s
4. Dynamic P icing Implemen a ion S a egies
AI enables sophis ica ed eal- ime p ice op imiza ion h ough se e al echnical app oaches ha ha e e olu ionized
how e aile s de e mine op imal p ice poin s. These sys ems p ocess immense olumes o ma ke da a o maximize
e enue while main aining compe i i e posi ioning and cus ome sa is ac ion.
Rule-based dynamic p icing ep esen s he ounda ion o au oma ed p icing sys ems, implemen ing algo i hmic logic
o adjus p ices based on p ede ined business pa ame e s. Acco ding o a sys ema ic li e a u e e iew o AI-based
dynamic p icing me hods, hese ule-based sys ems ypically ollow IF-THEN logic s uc u es ha espond o speci ic
ma ke condi ions like in en o y le els, ime-based ac o s, and compe i o mo emen s [7]. While concep ually
s aigh o wa d, hese sys ems ha e demons a ed signi ican business impac , wi h e-comme ce implemen a ions
epo ing e enue inc eases be ween 5-15% a e deploymen . The p ima y ad an age o ule-based app oaches lies in
hei explainabili y and manageable complexi y, making hem pa icula ly aluable o o ganiza ions beginning hei
dynamic p icing jou ney. Howe e , as no ed in he esea ch, hese sys ems ace limi a ions in cap u ing complex ma ke
dynamics and o en equi e subs an ial manual o e sigh o emain e ec i e in apidly changing ma ke s.
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 458-463
461
Machine lea ning-based p icing ep esen s a signi ican ad ancemen in p icing sophis ica ion, le e aging p edic i e
modeling o de e mine op imal p ice poin s. Recen esea ch on ML-d i en dynamic p icing s a egies in e-comme ce
iden i ies eg ession echniques, decision ees, andom o es s, and neu al ne wo ks as he mos commonly
implemen ed algo i hms o p edic ing p ice elas ici y and cus ome willingness o pay [8]. These sys ems can p ocess
hund eds o a iables simul aneously o iden i y p icing oppo uni ies ha would be impossible o disco e manually.
Acco ding o implemen a ion s udies, e-comme ce companies u ilizing ML-based p icing ha e achie ed e enue
imp o emen s o 2-5% and p o i inc eases o 3-8% compa ed o s a ic p icing app oaches [8]. Wha dis inguishes hese
sys ems is hei abili y o con inuously lea n om ma ke esponses and adap o changing condi ions wi hou explici
ep og amming.
Geo-p icing algo i hms ex end dynamic p icing by inco po a ing loca ion-speci ic ac o s in o p icing decisions. These
sys ems analyze egional demand pa e ns, local compe i ion, and demog aphic a iables o op imize p ices o speci ic
geog aphical a eas. As no ed in comp ehensi e e iews o AI p icing me hods, geo-p icing implemen a ions commonly
u ilize clus e ing algo i hms o iden i y simila ma ke egions be o e applying loca ion-speci ic p icing modi ie s [7].
Ube 's su ge p icing algo i hm exempli ies his app oach, analyzing eal- ime demand densi y ac oss mic o- egions and
implemen ing dynamic mul iplie s ha ha e inc eased d i e a ailabili y du ing peak demand pe iods while op imizing
e enue. The sys em p ocesses da a om millions o ides daily, adjus ing p ices in nea eal- ime based on
sophis ica ed demand o ecas ing models ha inco po a e wea he condi ions, a ic pa e ns, and special e en s [8].
Table 3 Al e na i e Ti le: Pe o mance Compa ison o AI-D i en Dynamic P icing S a egies [7, 8]
Dynamic
P icing
App oach
Key Pe o mance
Indica o
Pe o mance
Imp o emen
P ima y Ad an age
Key Technologies
Rule-based
Dynamic P icing
Re enue Inc ease
5-15%
Explainabili y and
manageable complexi y
IF-THEN logic s uc u es
Machine
Lea ning-based
P icing
Re enue
Imp o emen
2-5%
Con inuous lea ning
om ma ke esponses
Reg ession, decision ees,
andom o es s, neu al
ne wo ks
Machine
Lea ning-based
P icing
P o i Inc ease
3-8%
Adap a ion wi hou
explici ep og amming
P edic i e modeling
Geo-p icing
Algo i hms
D i e A ailabili y
(Ube case)
No speci ied
Loca ion-speci ic
op imiza ion
Clus e ing algo i hms,
demand o ecas ing
5. Con e sa ional AI and Cus ome Engagemen Sys ems
Ad anced Na u al Language P ocessing (NLP) and dialogue managemen sys ems powe AI-d i en cha bo s and i ual
assis an s ha ha e undamen ally ans o med cus ome se ice and engagemen in e ail en i onmen s. These
echnologies enable sophis ica ed in e ac ions ha closely mimic human con e sa ion while ope a ing a
unp eceden ed scale and e iciency.
Con e sa ional AI amewo ks ha e e ol ed d ama ically in ecen yea s, wi h mode n sys ems achie ing ema kable
sophis ica ion in unde s anding cus ome in en . Acco ding o a comp ehensi e e iew o con e sa ional AI-based
cha bo s, hese sys ems can now be classi ied in o ou dis inc ca ego ies: ule-based, e ie al-based, gene a i e, and
hyb id models - wi h each o e ing speci ic ad an ages o di e en e ail applica ions [9]. The mos ad anced
implemen a ions u ilize ans o me a chi ec u e and ine- uning app oaches o achie e con ex awa eness and
main ain cohe en mul i- u n con e sa ions. Resea ch indica es ha implemen a ion o hese sys ems in e ail
en i onmen s has educed cus ome se ice cos s by up o 30% while simul aneously imp o ing esponse imes by
80%. The echnological e olu ion has been pa icula ly no able in in en ecogni ion capabili ies, wi h mode n sys ems
demons a ing accu acy a es exceeding 90% ac oss di e se linguis ic exp essions compa ed o jus 60-70% i e yea s
ago [9].
Voice assis an in eg a ion ep esen s a apidly g owing channel o e ail engagemen , wi h sys ems like Amazon Alexa
and Google Assis an p ocessing millions o shopping- ela ed que ies daily. Acco ding o indus y analysis, oice
comme ce is p ojec ed o each $80 billion in global ma ke size by 2023, wi h app oxima ely 75% o U.S. households
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 458-463
462
expec ed o own a leas one sma speake [10]. These sys ems ha e ans o med how consume s in e ac wi h e ail
b ands, wi h 43% o sma speake owne s using hei de ices o shopping ac i i ies anging om p oduc esea ch o
di ec pu chasing. The echnology has shown pa icula s eng h in eo de scena ios, wi h oice-based eo de ing
accoun ing o 20% o all epea pu chases among egula oice assis an use s, demons a ing he o ma 's
con enience o habi ual pu chasing beha io [10].
In elligen cha bo a chi ec u es ha e become inc easingly sophis ica ed in hei abili y o handle complex cus ome
in e ac ions. Mode n e ail implemen a ions inco po a e p oduc ecommenda ion engines, na u al language
unde s anding, and pe sonaliza ion capabili ies o deli e highly ailo ed expe iences. Sepho a's Facebook Messenge
cha bo exempli ies his echnology, p o iding pe sonalized beau y ecommenda ions based on cus ome p e e ence
p o iles. The sys em engages cus ome s h ough in e ac i e ques ionnai es o unde s and skin ype, colo p e e ences,
and beau y goals be o e deli e ing highly ele an p oduc sugges ions. Acco ding o indus y epo s, implemen a ion
o such specialized e ail cha bo s has led o a e age inc eases o 25% in con e sion a es and 30% in cus ome
sa is ac ion sco es [9], while educing he need o human in e en ion in ou ine cus ome inqui ies by up o 80% [10].
6. Technical Implemen a ion o In-S o e Pe sonaliza ion
Physical e ail en i onmen s a e being ans o med h ough AI-powe ed echnologies ha b idge he gap be ween
digi al con enience and angible shopping expe iences. These inno a ions a e eshaping cus ome expec a ions while
p o iding e aile s wi h powe ul new ools o pe sonaliza ion and engagemen .
Sma mi o sys ems ep esen one o he mos isible implemen a ions o AI in physical e ail spaces. Acco ding o
esea ch published in Technological Fo ecas ing and Social Change, hese in e ac i e sys ems inco po a e ad anced
compu e ision algo i hms ha can analyze a cus ome 's body dimensions and s yle p e e ences o p o ide
pe sonalized ecommenda ions [11]. The echnology has e ol ed signi ican ly, wi h mode n sys ems able o ecognize
o e 100 body poin s and c ea e highly accu a e i ual ep esen a ions. S udies indica e ha s o es implemen ing
sma mi o s ha e seen cus ome engagemen inc ease by up o 59%, wi h a e age i ing oom dwell ime ex ending
om 10 minu es o 17 minu es. These sys ems e ec i ely add ess size unce ain y issues, which has been shown o
educe e u n a es by 18-23% in appa el e ail. The in eg a ion o hese echnologies ep esen s a signi ican
ad ancemen in wha esea che s call "phygi al e ail" - he meaning ul con e gence o physical and digi al shopping
expe iences ha enhances cus ome alue [11].
Facial ecogni ion deploymen s enable e aile s o iden i y e u ning cus ome s and deli e highly pe sonalized in-s o e
expe iences. Acco ding o comp ehensi e esea ch on AI-d i en pe sonaliza ion, acial ecogni ion sys ems in e ail
en i onmen s can now iden i y cus ome s wi h accu acy a es exceeding 97% unde no mal ligh ing condi ions [12].
When implemen ed esponsibly wi h p ope consen amewo ks, hese sys ems allow e aile s o ecognize high- alue
cus ome s and deli e ailo ed expe iences ha signi ican ly impac pu chase beha io . S udies show ha cus ome s
who ecei e pe sonalized in-s o e ecommenda ions based on hei pu chase his o y and p e e ences demons a e
31% highe sa is ac ion a es and 22% inc eased pu chase likelihood compa ed o hose ecei ing gene ic assis ance
[12].
Augmen ed eali y shopping applica ions ha e d ama ically expanded he possibili ies o p oduc isualiza ion and
in e ac ion. Resea ch indica es ha AR-enabled shopping expe iences inc ease consume pu chase in en ion by
c ea ing a " y be o e you buy" expe ience ha educes pu chase unce ain y [11]. In beau y e ail, i ual makeup y-
on applica ions ha e been shown o inc ease con e sion a es by 2.5 imes compa ed o adi ional shopping
expe iences. Simila ly, u ni u e e aile s implemen ing AR isualiza ion epo 40% highe cus ome con idence in
pu chase decisions. Nike's lagship s o e in New Yo k exempli ies hese echnologies h ough i s in eg a ion o sma
mi o s and digi al i ing ooms. This implemen a ion aligns wi h esea ch indings showing ha 80% o Gene a ion Z
shoppe s p e e e aile s ha o e AR expe iences, wi h 61% speci ically indica ing hey would choose hese e aile s
o e compe i o s lacking such echnology [12].
7. Conclusion
AI-d i en pe sonaliza ion has undamen ally ans o med he e ail landscape by enabling unp eceden ed le els o
cus ome unde s anding and engagemen . F om sophis ica ed ecommenda ion engines ha an icipa e consume
needs o dynamic p icing sys ems ha op imize alue o bo h e aile s and cus ome s, hese echnologies ha e become
essen ial componen s o compe i i e e ail s a egy. Con e sa ional AI has ede ined cus ome se ice while in-s o e
pe sonaliza ion echnologies ha e e i alized physical e ail by c ea ing seamless connec ions be ween digi al
Wo ld Jou nal o Ad anced Resea ch and Re iews, 2025, 26(02), 458-463
463
con enience and angible shopping expe iences. As hese echnologies con inue o e ol e, e aile s who hough ully
implemen AI pe sonaliza ion solu ions s and o gain signi ican ad an ages in cus ome acquisi ion, e en ion, and
li e ime alue maximiza ion. The u u e o e ail pe sonaliza ion lies in c ea ing in elligen , adap i e, and e hically sound
sys ems ha espec cus ome p i acy while deli e ing inc easingly meaning ul and ele an expe iences.
Re e ences
[1] Sha haboina Raju, "The impac o A i icial in elligence on enhancing he in-s o e cus ome expe ience in e ail
sec o ," Jou nal o Managemen Resea ch and Analysis, 2025. [Online]. A ailable:
h ps://pd .ipinno a i e.com/pd /24113
[2] Tanawa Teepapal, "AI-d i en pe sonaliza ion: Un a eling consume pe cep ions in social media engagemen ,"
Compu e s in Human Beha io , 2025. [Online]. A ailable:
h ps://www.sciencedi ec .com/science/a icle/abs/pii/S0747563224004175
[3] Jimmy Newson, "Mas e ing Ma ke Segmen a ion wi h AI Insigh s," Mo ing Fo wa d Small Business, Ap il 2024.
[Online]. A ailable: h ps://mo ing o wa dsmallbusiness.com/segmen a ion-wi h-ai-expe -ma ke -insigh s/
[4] Zeinab Shahbazi e al., "Towa d Imp o ing he P edic ion Accu acy o P oduc Recommenda ion Sys em Using
Ex eme G adien Boos ing and Encoding App oaches," Symme y, 2020. [Online]. A ailable:
h ps://www.mdpi.com/2073-8994/12/9/1566
[5] Hongde Zhou, Fei Xiong and Hongshu Chen, "A Comp ehensi e Su ey o Recommende Sys ems Based on Deep
Lea ning," Applied Sciences, 2023. [Online]. A ailable: h ps://www.mdpi.com/2076-3417/13/20/11378
[6] Rim Fakh akh, Ben amma Anis and Chok i Ben Ama , "Deep Lea ning-Based Recommenda ion: Cu en Issues
and Challenges," In e na ional Jou nal o Ad anced Compu e Science and Applica ions, 2017. [Online]. A ailable:
h ps://www. esea chga e.ne /publica ion/322248919_Deep_Lea ning-
Based_Recommenda ion_Cu en _Issues_and_Challenges
[7] Ch is oph Tomi za, Ul i Ib ahimli and Lukas-Valen in He m, "AI-Based Me hods o Dynamic P icing in E-
Comme ce: A Sys ema iza ion o Li e a u e," The 32nd Eu opean Con e ence on In o ma ion Sys ems, 2024.
[Online]. A ailable: h ps://www. esea chga e.ne /publica ion/381011158_AI-
Based_Me hods_o _Dynamic_P icing_in_E-Comme ce_A_Sys ema iza ion_o _Li e a u e
[8] Raouya El Youbi, Fayçal Messaoudi and Manal Loukili, "Machine Lea ning-d i en Dynamic P icing S a egies in
E-Comme ce," 14 h In e na ional Con e ence on In o ma ion and Communica ion Sys ems (ICICS), 2023.
[Online]. A ailable: h ps://www. esea chga e.ne /publica ion/376222315_Machine_Lea ning-
d i en_Dynamic_P icing_S a egies_in_E-Comme ce
[9] M. Vishal and H. Vishalakshi P abhu, "A Comp ehensi e Re iew o Con e sa ional AI-Based Cha bo s: Types,
Applica ions, and Fu u e T ends," In e ne o Things (IoT): Key Digi al T ends Shaping he Fu u e, 2023. [Online].
A ailable:
h ps://www. esea chga e.ne /publica ion/372601748_A_Comp ehensi e_Re iew_o _Con e sa ional_AI-
Based_Cha bo s_Types_Applica ions_and_Fu u e_T ends
[10] O ange Man a, "How can business p epa e o he ise o he u u e o oice o comme ce in e-comme s," O ange
Man a, Ma . 2024. [Online]. A ailable: h ps://www.o angeman a.com/blog/ u u e-o - oice-comme ce-in-
ecomme ce/
[11] Ayo unde Ogunjimi e al., "Sma mi o ashion echnology o he e ail chain ans o ma ion," Technological
Fo ecas ing and Social Change, 2021. [Online]. A ailable:
h ps://www.sciencedi ec .com/science/a icle/abs/pii/S0040162521005515
[12] Falope Samson, "AI-D i en Pe sonaliza ion In E-Comme ce," Resea chGa e, 2025. [Online]. A ailable:
h ps://www. esea chga e.ne /publica ion/389626209_AI-DRIVEN_PERSONALIZATION_IN_E-COMMERCE
