Ho mann, Pe e ; U bach, Nils; Lanzl, Julia; Desouza, Ke in C.
A icle — Published Ve sion
AI-enabled in o ma ion sys ems: Teaming up wi h
in elligen agen s in ne wo ked business
Elec onic Ma ke s
P o ided in Coope a ion wi h:
Sp inge Na u e
Sugges ed Ci a ion: Ho mann, Pe e ; U bach, Nils; Lanzl, Julia; Desouza, Ke in C. (2024) : AI-enabled
in o ma ion sys ems: Teaming up wi h in elligen agen s in ne wo ked business, Elec onic Ma ke s,
ISSN 1422-8890, Sp inge , Be lin, Heidelbe g, Vol. 34, Iss. 1,
h ps://doi.o g/10.1007/s12525-024-00734-y
This Ve sion is a ailable a :
h ps://hdl.handle.ne /10419/315705
S anda d-Nu zungsbedingungen:
Die Dokumen e au EconS o dü en zu eigenen wissenscha lichen
Zwecken und zum P i a geb auch gespeiche und kopie we den.
Sie dü en die Dokumen e nich ü ö en liche ode komme zielle
Zwecke e iel äl igen, ö en lich auss ellen, ö en lich zugänglich
machen, e eiben ode ande wei ig nu zen.
So e n die Ve asse die Dokumen e un e Open-Con en -Lizenzen
(insbesonde e CC-Lizenzen) zu Ve ügung ges ell haben soll en,
gel en abweichend on diesen Nu zungsbedingungen die in de do
genann en Lizenz gewäh en Nu zungs ech e.
Te ms o use:
Documen s in EconS o may be sa ed and copied o you pe sonal
and schola ly pu poses.
You a e no o copy documen s o public o comme cial pu poses, o
exhibi he documen s publicly, o make hem publicly a ailable on he
in e ne , o o dis ibu e o o he wise use he documen s in public.
I he documen s ha e been made a ailable unde an Open Con en
Licence (especially C ea i e Commons Licences), you may exe cise
u he usage igh s as speci ied in he indica ed licence.
h p://c ea i ecommons.o g/licenses/by/4.0/
Vol.:(0123456789)
Elec onic Ma ke s (2024) 34:52
h ps://doi.o g/10.1007/s12525-024-00734-y
PREFACE
AI-enabled in o ma ion sys ems: Teaming upwi hin elligen agen s
inne wo ked business
Pe e Ho mann1,2· NilsU bach1,3,6· JuliaLanzl1,4,6· Ke inC.Desouza5
Published online: 4 Oc obe 2024
© The Au ho (s) 2024
In oduc ion
A i icial in elligence (AI) echnologies p omise new ways
o sol e (exis ing) p oblems and o c ea e inno a ions, esul -
ing in new pa hs owa d business alue. In his way, no el
ans o ma ions a e on he ho izon o ealize oppo uni ies
we ha e no ye concei ed (o hough a e cu en ly easible).
While he unde s anding o wha cons i u es AI echnolo-
gies has con inued o change o e ime, machine lea ning
(Janiesch e al., 2021) and—mo e ecen ly—la ge language
models, espec i ely ounda ion models, ha e become he
ocus o applied esea ch and p ac ice (Banh & S obel,
2023). Besides AI applica ions’ capabili ies o ou pe o m
humans in ce ain asks, i is he “abili y o lea n and ac
au onomously [ ha ] makes in elligen echnological ac o s
e y di e en om mos echnologies his o ically used in
o ganiza ions” (Bailey e al., 2019, p.643).
The as de elopmen and mani old oppo uni ies o AI
applica ions mo i a e esea ch o gain a deepe unde s anding
o AI-enabled in o ma ion sys ems, pa icula ly ega ding he
ole o in elligen agen s and hei impac s on ne wo ked busi-
nesses. Fo his edi o ial, we unde s and in elligen agen s as
“a compu e sys em ha is capable o lexible au onomous
ac ion in o de o mee i s design objec i es” (Jennings &
Woold idge, 1998, p.4). Howe e , we a gue ha he p ope y
o lexibili y and au onomy should no be implici ly assumed
as gi en and ha ou discou se bene i s om dis inguish-
ing au oma ed, semi-au onomous, and au onomous sys ems.
Hence, we ecommend a non-bina y unde s anding o au on-
omy by desc ibing he echnical a i ac ’s sel -su iciency
(o au a ky), i s cons ain s o independence in ul illing i s
goals, and i s ne wo king capabili ies.
In line wi h Be en e e al. (2021) and Bai d and Ma up-
ing (2021), we alue he ichness o he agen pe spec i e
o esea ching con empo a y and u u e socio- echnical
phenomena. The agen pe spec i e is calling in o ques ion
ou exis ing assump ions abou he in eg a ion o AI ech-
nologies in wo k sys ems. Beyond in elligen agen s’ capa-
bili ies o con ibu e o wo k sys ems, insc u abili y issues
as echnology-implied cons ain s a e pa icula ly salien in
oday’s esea ch discussions (Be en e e al., 2021).
This opical collec ion ocuses on he app op ia e design
o AI-enabled in o ma ion sys ems (IS), hei accompany-
ing managemen , and he ans o ma ional p ocesses. This
comes wi h mul i ace ed and ascina ing ques ions o he IS
discou se whose answe s ake a socio- echnical pe spec i e
on he changing in e ac ion wi hin o ganiza ions (esp. indi-
iduals, eams) and be ween hem. Pa icula ly, ou ocus is
on o ms o ne wo ked business whe e in elligen and human
agen s in e ac o economic pu poses wi hin one o among
mul iple ie s in economic alue chains.
This a icle is o ganized as ollows: In he nex sec ion,
we ou line cen al issues and opics in he ield o AI-enabled
in o ma ion sys ems s uc u ed a ound he ecological wo k
sys ems amewo k ha we in oduce o ha pu pose. In
he ollowing sec ion, we p esen ou pe spec i e on u u e
esea ch a enues by discussing i e majo AI ans o ma ion
headwinds ha we de i ed om ou p ac ical expe iences
* Pe e Ho mann
pe e .ho mann@uni-bay eu h.de
* Nils U bach
[email p o ec ed]
1 FIM Resea ch Cen e , Uni e si y o Bay eu h, Bay eu h,
Ge many
2 appliedAI Ini ia i e GmbH, Munich, Ge many
3 Resea ch Lab o Digi al Inno a ion & T ans o ma ion,
F ank u Uni e si y o Applied Sciences, F ank u ,
Ge many
4 Chai o Digi al Managemen , Uni e si y o Hohenheim,
S u ga , Ge many
5 School o Managemen , Facul y o Business & Law,
Queensland Uni e si y o Technology, B isbane, Aus alia
6 B anch Business andIn o ma ion Sys ems Enginee ing,
F aunho e Ins i u e o Applied In o ma ion Technology
FIT, Augsbu g, Ge many
Elec onic Ma ke s (2024) 34:5252 Page 2 o 8
and p ac i ione eedback om ou ne wo k. In he inal sec-
ion, we p o ide an o e iew o he accep ed con ibu ions
ha we included in he ini ial compila ion o his opical
collec ion.
Cen al issues and opics
We bo owed he concep ualiza ion o he sys ems su ound-
ing he indi idual om ecological sys ems heo y (B on-
enb enne , 1979). Fi s , his allows us o adop a human-
cen e ed pe spec i e on AI-enabled in o ma ion sys ems,
os e ing empa hic design (Leona d-Ba on, 1995). Second,
i enables s udying he ela ionship be ween (in elligen and)
human agen s wi hin hei en i onmen . Thi d, i enables
us o classi y exis ing esea ch s ands and de i e u u e
esea ch oppo uni ies.
The ecological wo k sys ems amewo k (see Fig.1)
consis s o i e le els: human agen , mic osys em, meso-
sys em, ecosys em, and mac osys em. A he cen e o he
amewo k is he human agen , acknowledging he necessi y
o a human-cen e ed pe spec i e on AI. The mic osys em
comp ises he in e ac ion o he human agen wi h in el-
ligen agen s and o he human agen s o accomplish asks
in bila e al collabo a ion. The mesosys em ep esen s he
in e connec ion be ween di e en mic osys ems (i.e., g oup
wo k in mul ila e al collabo a ion). The exosys em cha ac-
e izes he links be ween social se ings ha do no in ol e
he human (i.e., dis an wo k, o example, wi h o he depa -
men s o companies). Finally, he mac osys em s ands o he
su ounding ecosys em ela ing o he sys em o ne wo ked
business, o en di ec ed h ough co po a e alues, goals, and
di ec i es.
Al hough he imp ession may a ise ha his amewo k
has a pu ely company-in e nal ocus, in e ac ions can also
in ol e ex e nal pa ies a each le el, be i cus ome in e ac-
ion, join de elopmen p ojec s, o gig wo ke s. Thus, we
emphasize he pe meabili y o co po a e bounda ies in he
ne wo ked business and, hus, he wo k sys em wi hin he
whole ecosys em.
Human agen
The amewo k’s cen e embodies ou human-cen e ed
app oach, aking in o accoun he posi i e and nega i e
consequences o using AI-enabled in o ma ion sys ems o
he human agen as well as he necessa y esou ces ha he
human agen needs o be p o ided. On he one hand, he e
can be many posi i e ou comes o he human agen , such
as elie om exhaus ing wo k o lea ning om in elligen
agen s. While consul ancies a e ou doing each o he wi h
business po en ial es ima es, we also wan o encou age
Fig. 1 Ecological wo k sys ems
amewo k
Human
agen
In elligen
agen s
Human
agen s
Tasks
Mic osys em
(bila e al collabo a ion)
Mesosys em
(mul ila e al collabo a ion)
Exosys em
(ou o he indi idual’s immedia e each)
Mac osys em
(socie y)
Co po a e
alues,
goals, and
di ec i es
G oup wo k
Dis an wo k
Co po a e wo k
ou come
Dis an
agen s
Elec onic Ma ke s (2024) 34:52 Page 3 o 8 52
conside ing how AI applica ions could imp o e wo k- ela ed
ai ness and heal h issues (e.g., compensa e o inequali ies
a wo k). On he o he hand, he e can also be ad e se ou -
comes (e.g., inc eased isk o digi al s ess) esul ing om,
among o he s, he eeling o job insecu i y o he excessi e
demands o con olling o unde s anding AI applica ion
ou comes. Howe e , he causes o s ess om AI-enabled
in o ma ion sys ems ha e no ye been comple ely unco -
e ed. Among o he s, we mo i a e esea ching he empo al
hy hm and o m o AI-enabled wo k sys ems and hei con-
sequences o human agen s (O likowski & Ya es, 2002). In
o de o ealize he p omised p oduc i i y gains, he asks
and hus also he skill equi emen s a e expec ed o change,
p obably as e han wi h he p e ious gene al-pu pose
echnologies (McA ee, 2024). I is o g ea impo ance ha
human agen s a e adequa ely p epa ed o deal wi h in el-
ligen agen s by ollowing ole- and skill-speci ic lea ne
pa hs.
Mic osys em
Con inuing wi h he second laye o he ecological wo k
sys ems amewo k, we ocus on he mic osys em ha
comp ises he in e ac ion o he human agen wi h in elli-
gen agen s and o he human agen s o accomplish asks in
bila e al collabo a ion. We ollow Jakob e al. (2024) and
desc ibe he ou essen ial ac i i ies o collabo a ion as ol-
lows: Fi s ly, (human and in elligen ) agen s need o com-
munica e in o de o es ablish goals, engage in nego ia ions
o de e mine he cou se o ac ion, and assess p og ess and
ou comes (Ma essich & Monsey, 1992; Te een, 1995).
Secondly, collabo a ion equi es agen s o p ocess asks,
meaning hey mus ake ac ion and join ly manage hei asks
o accomplish sha ed objec i es (Wang e al., 2020). Thi dly,
agen s mus de e mine which ac ions will be unde aken by
each pa icipan and alloca e esponsibili ies o speci ic
asks (Te een, 1995). Finally, collabo a ion in ol es agen s
nego ia ing and deciding on he le el o con ol (i.e., au ho -
i y) each pa icipan will ha e abou hei ac ions, ensu ing
coo dina ed ope a ions (Te een, 1995).
The ypes o human-AI ela ionships can be mani old
wi hin he spec um o au oma ing and augmen ing human
wo k. Mölle s e al. (2024) o ganize he design space by
dis inguishing he mos p ominen human-AI ela ionships
a he wo kplace in o decision suppo , AI in he loop, algo-
i hmic managemen , human-AI collabo a ion and eams,
human in he loop, and ull delega ion o AI. In elligen
agen s may no only pe o m asks bu unc ion as en i ies
guiding human ac ions and may e en al e he dynamics o
ask delega ion be ween humans and in elligen agen s. By
his, we do no mean ha in elligen agen s achie e com-
ple e au onomy bu ha humans use in elligen agen s as
ools equipped wi h delega ion capabili ies and au ho i y.
Fu u e in elligen agen s migh e en blu he lines be ween
pa icipan s and echnological componen s wi hin a wo k
sys em by e e sing he low o wo k delega ion o elimina -
ing humans om hese sys ems (Bai d & Ma uping, 2021).
Resea ch has al eady demons a ed ha human-AI sys ems
can—in some cases—yield supe io esul s when he in el-
ligen agen akes a leading ole, delega ing asks o humans
a he han he e e se (e.g., Leibig e al., 2022). This del-
ega ion e e sal holds subs an ial implica ions o human-AI
collabo a ion (Benbya e al., 2020; Wesche & Sonde egge ,
2019). Fo ins ance, he s udy conduc ed by Guggenbe ge
e al. (2023) del es in o his phenomenon using he heo e i-
cal amewo k o p incipal-agen heo y. I iden i ies no el
sou ces o ension a ising speci ically in AI- o-human del-
ega ion, emphasizing he need o specialized mechanisms
o add ess he ensuing challenges.
Mesosys em
The hi d laye o he ecological wo k sys ems amewo k,
he mesosys em, ep esen s he in e connec ion be ween di -
e en mic osys ems, such as g oup wo k in mul ila e al col-
labo a ion. So a , human–compu e in e ac ion li e a u e has
concen a ed on indi idual in e ac ions, lea ing ou a wo k
sys em pe spec i e on he mul ila e al in e ac ions be ween
human and in elligen agen s when p ocessing and delega -
ing asks as well as de ining au ho i ies and esponsibili ies
(Jakob e al., 2024). Building on Hinsen e al. (2022), Jakob
e al. (2024) in oduce a amewo k ha acili a es desc ib-
ing and analyzing wo k sys ems o human and in elligen
agen s’ collabo a ion beyond bila e al in e ac ion.
One p ima y complica ion wi hin his pe spec i e is he
e ec i e coo dina ion and managemen o asks among
di e se eam membe s (bo h human and in elligen agen s).
This is exempli ied in he con ex o so wa e enginee s
wo king alongside AI copilo s like Gi Hub Copilo . These
collabo a ion se ings necessi a e he in eg a ion o mul i-
ple human and in elligen agen s’ con ibu ions in o la ge
so wa e p ojec s. Ensu ing seamless collabo a ion equi es
clea ly de ined oles, p ocesses, and an unde s anding o
he mu ual s eng hs and limi a ions. Fo ins ance, while
AI copilo s can signi ican ly enhance coding e iciency and
e o de ec ion, human enginee s mus o e see and alida e
AI-gene a ed ou pu s o main ain quali y and cohe ence. On
pape , human supe ision is easy o es ablish, bu he e is
a isk ha he con ol e ec will ail o ma e ialize in day-
o-day p ac ice.
O e all, he mesosys em laye unde sco es he impo ance
o s a egically designed collabo a ion be ween human and
in elligen agen s, highligh ing he need o obus ame-
wo ks o manage hese complex, mul ila e al wo k se ings.
P ac ice will be con on ed wi h new ques ions ega ding
he design and coo dina ion o g oup wo k o le e age he
Elec onic Ma ke s (2024) 34:5252 Page 4 o 8
complemen a i ies o human and in elligen agen s’ capabili-
ies wi hou o e bu dening humans. No leas , p ac i ione s
a e inc easingly con on ed wi h ma e s o p ocess edesign.
Exosys em
The ou h laye o he ecological wo k sys ems amewo k is
he exosys em, which cha ac e izes he connec ions be ween
social se ings ha do no di ec ly in ol e he indi idual.
I includes dis an wo k ela ionships, such as hose wi h
o he depa men s o ex e nal companies. The exosys em
is an impo an laye ha explains why syne gies be ween
mic osys ems a e le un apped o unin ended consequences
a ise. This is especially c ucial since he b oad ele ance o
AI can lead o nume ous, o en uncoo dina ed, ini ia i es.
In p ac ice, companies a e se ing up o p o essionalizing
hei AI ope a ing model o align he mic osys ems acco d-
ing o he co po a e (AI) s a egy and e hical p inciples and
alues. Lessons om digi aliza ion li e a u e, such as he
“shadow IT” issue, whe e di e en pa s o an o ganiza ion
implemen hei own solu ions wi hou cen alized o e sigh
(Fue s enau & Ro he, 2014), p o ide insigh s, bu i is also
c ucial o ecognize he AI-speci ic aspec s. In pa icula ,
he insc u abili y ace o AI (i.e., being unin elligible o
mul iple audiences) (Be en e e al., 2021) and he limi ed
obus ness o p obabilis ic ou comes lead o new go e nance
challenges. Lämme mann e al. (2024) highligh ha man-
aging AI applica ions e ec i ely equi es a obus in o ma-
ion exchange among di e se s akeholde s. Wi hou adequa e
in o ma ion p ocessing, ask unce ain y ises, unde mining
AI ope a ions. O ganiza ions can be e manage AI applica-
ions by os e ing an en i onmen o anspa en and e icien
in o ma ion exchange, hus minimizing ope a ional unce -
ain ies and unin ended consequences and maximizing hei
po en ial bene i s. The eby, e hical ques ions will be com-
monplace when de eloping o applying AI echnologies
(e en i e hical p oblems do no always a ise depending on
he con ex ). The managemen challenge lies in iden i ying
and unde s anding e hically p oblema ic ques ions ea ly and
gi ing answe s aligned wi h he co po a e e hical p inciples
and alues.
Mac osys em
Finally, he ou e laye o he ecological wo k sys ems
amewo k, he mac osys em, ep esen s he su ounding
ecosys em o ne wo ked businesses, which is o en guided
by co po a e alues, goals, and di ec i es. AI should no be
pu sued me ely o i s own sake bu o ein o ce he o gani-
za ional iden i y and deli e angible business alue. Wi hin
his pe spec i e, companies ace he signi ican challenge o
keeping pace wi h apid echnological ad ancemen s while
cap u ing enough alue o ecoup hei in es men s. Balanc-
ing inno a ion wi h sus ainable business p ac ices equi es
ca e ul na iga ion o ex e nal p essu es and in e nal di ec-
i es. As o ganiza ions s i e o s ay compe i i e, aligning
hei mac osys em s a egies wi h hei o e a ching co po-
a e ision becomes c ucial. The challenge, he e o e, is o
ansla e he echnology-d i en momen um—among o he s
esul ing om he ea o missing ou (FOMO) a he man-
agemen le el—in o p oblem-sol ing AI applica ions. To
aise business alue, o ganiza ions should c i ically e lec
on he business alue po en ial o AI use cases ea ly on and
in es in he us wo hiness o AI-enabled in o ma ion sys-
ems. To ul ill he c i e ion o us wo hiness, AI-enabled
in o ma ion sys ems should be law ul, e hically aligned, and
obus om a echnical and social pe spec i e (Eu opean
Commission 2019).
A dedica ed AI s a egy acili a es na iga ing he com-
plexi ies o he AI ans o ma ion ha comp ises he iden-
i ica ion and ealiza ion o AI use cases as well as he
enhancemen o he o ganiza ion’s AI ma u i y (i.e., capa-
bili ies o iden i y and ealize u u e AI use cases e ec i ely
and e icien ly). Besides de ining he s a egic a ge s and AI
applica ion ields, he AI s a egy should o mula e e hical
p inciples and alues ega ding he de elopmen , ope a ion,
and use o AI applica ions. Howe e , his unde aking is
no i ial, as isks o en canno be uled ou , so o ganiza-
ions mus speci y how good is good enough (e.g., pe o -
mance h esholds). Inc easing an o ganiza ion’s AI ma u-
i y equi es no only in es men s in echnology bu also
o ganiza ional capabili ies and complemen a y asse s (Be g
e al., 2023; Duda e al., 2024; Jöhnk e al., 2021). Wi h-
ou a comp ehensi e unde s anding o ele an esou ces
and hei impac s on de eloping, ope a ing, and using AI-
enabled in o ma ion sys ems, o ganiza ions isk ine icien
esou ce alloca ion and o e seeing esou ce dependencies
(Duda e al., 2024). Wi h AI being a “mo ing on ie o
bo h inc easing pe o mance and inc easing scope” (Be -
en e e al., 2021), an o ganiza ion’s AI ma u i y can also
decline wi hou main aining a con inuous ans o ma ion.
Accele a ing he AI ans o ma ion mus be balanced wi h
s a egic alignmen o ensu e cohe en p og ess ac oss he
o ganiza ion. Insigh s om p e ious s a egy esea ch, such
as digi al s a egy, highligh he impo ance o adap abili y
and ongoing e alua ion, ensu ing ha AI ini ia i es a e bo h
inno a i e and s a egically g ounded.
Fu u e esea ch pe spec i es
To c ea e sus ainable alue h ough AI applica ions, o gani-
za ions mus apply AI echnologies pu pose ully and plan
and ca y ou o ganiza ional ini ia i es ha inc ease he
Elec onic Ma ke s (2024) 34:52 Page 5 o 8 52
o ganiza ion’s AI ma u i y. F om ou p ac ical expe iences
and p ac i ione eedback, we iden i ied i e signi ican AI
ans o ma ion headwinds:
1. Complexi y o coo dina ing ini ia i es (e.g., due o he
b oad scope o AI po en ials and in ol ed s akeholde s)
2. Lack o o ien a ion (e.g., due o he una ailabili y o
bluep in s o highe ma u i y le els)
3. Limi ed a sigh in a as -paced echnology en i onmen
(e.g., ega ding AI applica ions’ u u e capabili ies and
egula ion)
4. O ganiza ional pa alysis in execu ing he AI s a egy
(e.g., due o PoC-pa alysis o poo ly o ches a ed in es -
men s)
5. Incompa ibili y o adi ional KPIs and a ge -se -
ing app oaches (e.g., due o unce ain y and isks in
Machine Lea ning p ojec s)
To coun e he AI ans o ma ion headwinds, o ganiza-
ions bene i om an AI ans o ma ion managemen ha
b idges he sys em’s bounda ies and, hus, ensu es he pe -
mea ion om s a egic conside a ions, such as an o ganiza-
ion’s AI ambi ion, o he indi idual employees and ice
e sa (c. . Fig.2).
In he ollowing sec ions, we ake a close look a h ee
pe mea ion issues ha we conside o be impo an . A he
same ime, we wan o wa n abou jus pu ing old wine in
new bo les. Ou expe ience sugges s ha o ganiza ional
esea ch indings in he AI ield a e o en no necessa ily
exclusi e o he AI ield. F om ou poin o iew, his could
also be a s eng h as long as he pape clea ly de ines wha
cons i u es he objec o esea ch and does no hide unde
he AI umb ella.
An a chi ec u e pe spec i e onhuman‑AI
collabo a ion inwo k sys ems
Resea ch in o human–compu e o human–machine in e -
ac ions has a long adi ion. Due o he p e ailing ocus
on bila e al in e ac ions, a no able gap exis s in guidance
ega ding he holis ic design o collabo a i e amewo ks
o human and in elligen agen s wi hin wo k sys ems. By
holis ic design, we e e o in eg a ing unc ional, economic,
ecological, and social conside a ions. We expec his guid-
ance o become mo e impo an wi h he p oli e a ion o AI
applica ions and, hus, ad oca e an a chi ec u al pe spec i e
o designing wo k sys ems ha can e ec i ely exploi he
collabo a i e po en ial be ween di e se agen s and a oid
se e e consequences.
Conside ing he collabo a i e po en ial, we look o -
wa d o lea ning mo e abou he complemen a y capabili-
ies o he di e en ac o s beyond he bina y classi ica ion
o human and so wa e agen s. We long o esea ch ha
does no p i ilege he p i ileged. Fo ins ance, his could
include esea ch on AI-enabled wo k sys ems wi h neu odi-
e se ac o s o he in eg a ion o employees wi h disabili ies
(Maddali e al., 2022). We also conside IS esea ch espon-
sible o iden i ying he nega i e consequences o inapp o-
p ia ely designed wo k sys ems o ou plane and i s li ing
beings. Explo a i e esea ch disco e ing hem could ocus
on an o ganiza ion, i s ne wo ked business, o he big pic u e
o ou socie y. An exempla y conce n is he uncon olled
cascading o bugs, low-quali y da a, alse o misin e p e ed
ou pu , o biases.
The ope a i e glue be ween hesys ems
While an app op ia e design o wo k sys ems is a necessa y
i s s ep, one should also accoun o i s execu ion. F om
ou p ac ical insigh s in o co po a e p ac ice, we in oduce
ou ypes o ope a i e glue ha can go hand in hand in hei
implemen a ion: echnical glue, p ocess glue, in o ma ion,
and social glue. Technical glue in he o m o glue o eus-
able code allows o he in eg a ion o ools and esou ces
(Duda e al., 2024). P ocess glue comp ises de elopmen ,
ope a ion, and go e nance p ocesses ha guide ac ions
(e.g., he Hou glass Model o O ganiza ional AI Go e n-
ance in oduced by Män ymäki e al. (2022)). In o ma ion
glue esul s om he sa is ac ion o he ac o s’ indi idual
in o ma ion needs so ha hey can ul ill hei esponsibili-
ies (Lämme mann e al., 2024). Social glue esul s om he
sense o social oge he ness in luencing he ac o ’s beha io .
Fo u u e esea ch, we assume exci ing ques ions a ound
he (alleged) ension be ween con ol and expe imen a ion.
Taking he machine lea ning li ecycle’s expe imen al na u e
in o accoun , equi emen s such as aceabili y migh be con-
side ed a slowing ac o . Howe e , lineage o expe imen
acking ools migh also ca alyze expe imen a ion. Thus,
esea ch on he use o ope a i e glue o he complian indus-
ializa ion o expe imen a ion seems p omising. Mo eo e ,
in ligh o he EU AI Ac , we see he need o esea ch how
o app oach he ope a i e glue so ha o ganiza ions can
Fig. 2 B idging he bounda ies
h ough AI ans o ma ion
managemen
Elec onic Ma ke s (2024) 34:5252 Page 6 o 8
e ec i ely and e icien ly ensu e compliance wi h egula ion
equi emen s. Conside ing he social glue, we ask ou sel es
wha cha ac e izes beha io in sys ems in which he p opo -
ion o human wo k is educed o he (pe cei ed) dis ance
be ween people is inc eased. In addi ion, we a e exci ed o
see how o ganiza ions can ensu e e hical alignmen ega d-
ing philosophically challenging ques ions, in eg a ing indi-
idual employee alues and co po a e policies. In his con-
ex , we also encou age o hink abou he b oade cons uc
o public alue and how IS esea ch can ge ahead o keep
pa wi h examining he social (public) alue o AI (Desouza
& Dawson, 2023).
Dynamically s ee ing heAI ans o ma ion
Since he AI ield is cons an ly changing, companies will be
challenged epea edly o keep pace as echnological change
can easily ou pace he possibili ies o o ganiza ional change.
While big ech companies o highly unded s a ups migh
be able o pa icipa e in hese aces, he digi al so e eign y
o he emaining companies is a isk. Howe e , he e is no
ans o ma ion bluep in ha can be gene alized o all com-
panies. Fo ins ance, he e a e slow indus ies, also called
asse -in ensi e indus ies, whe e change o digi al echnolo-
gies is slow due o long p ojec ime ames, ega dless o
how exci ing inno a ion po en ials migh be (Buck e al.,
2023). Thus, we deem i essen ial o ind ways ha enable
all companies o su he AI wa es so e eignly.
We encou age esea ch on how o dynamically s ee he
AI ans o ma ion, in eg a ing all sys ems o he ecological
wo k sys ems amewo k. We highligh h ee p essing issues
om ou p ac ical expe ience and p ac i ione eedback: (1)
Wha a e e ec i e, measu able objec i es o key pe o -
mance indica o s o he AI ans o ma ion, and how can
hey be ope a ionalized ac oss he o ganiza ion in a s ee ing
me hod? (2) How can o ganiza ions design obus o adap-
i e alue c ea ion and cap u e mechanisms? (3) How can
con inuous and di e se employee de elopmen beyond s a ic
lea ning pa hs be e icien ly and e ec i ely app oached?
Accep ed pape s
We ha e accep ed ou pape s o inclusion in he ini ial
compila ion o his opical collec ion. Each a icle explo es
di e en aspec s o he special sec ion’s ocus on AI-enabled
in o ma ion sys ems.
The i s a icle in his opical collec ion, In o ma ion
P o ision Measu es o Voice Agen P oduc Recommenda-
ions – The E ec o P ocess Explana ions and P ocess Visu-
aliza ions on Fai ness Pe cep ions, by Helena Wei h and
Ch is ian Ma (Wei h & Ma , 2023), examines he impac
o addi ional in o ma ion measu es on use s’ pe cep ions o
ai ness and hei beha io al esponses o oice agen p od-
uc ecommenda ions (VAPRs). Due o inhe en opaci ies
in AI ecommenda ion engines and he limi a ions o audio-
based communica ion, use s may eel un ai ly ea ed du ing
hei pu chase decisions, po en ially ha ming e aile s. The
au ho s u ilize in o ma ion p ocessing and s imulus-o gan-
ism- esponse heo y o explo e how p ocess explana ions
and p ocess isualiza ions in luence use s’ ai ness pe cep-
ions and beha io s. Th ough wo expe imen al s udies, hey
disco e ed ha p ocess explana ions enhance use s’ sense
o ai ness, whe eas p ocess isualiza ions do no ha e he
same e ec . The s udy highligh s ha explana ions ailo ed
o use s’ p o iles and pas pu chase beha io s e ec i ely
imp o e ai ness pe cep ions. This esea ch ad ances he
li e a u e on ai and explainable AI by add essing audio-
based cons ain s in VAPRs and linking hese ac o s o use
pe cep ions and eac ions. The indings p o ide aluable
insigh s o p ac i ione s on employing in o ma ion p o i-
sion measu es o mi iga e pe cep ions o un ai ness and p e-
en nega i e cus ome beha io s.
The second a icle in he opical collec ion, AI Li e acy
o he Top Managemen : An Uppe Echelons Pe spec i e
on Co po a e AI O ien a ion and Implemen a ion Abili y,
by Ma c Pinski, Thomas Ho mann, and Alexande Benlian
(Pinski e al., 2024), explo es he in luence o op manage-
men eam (TMT) AI li e acy on a i m’s abili y o gene -
a e alue h ough AI, ocusing on wo key cha ac e is ics:
AI o ien a ion and AI implemen a ion abili y. G ounded in
uppe echelons heo y, he s udy in es iga es how he AI
knowledge o a i m’s TMT a ec s i s capaci y o iden i y
AI oppo uni ies (AI o ien a ion) and o execu e AI ini ia-
i es (AI implemen a ion abili y). The au ho s also conside
he mode a ing ole o i m ype, dis inguishing be ween
s a ups and incumben i ms. Using obse a ional da a
om 6986 execu i es’ LinkedIn p o iles and i m da a om
10-k s a emen s, he s udy inds ha TMT AI li e acy sig-
ni ican ly enhances bo h AI o ien a ion and implemen a ion
abili y. Mo eo e , AI o ien a ion media es he ela ionship
be ween TMT AI li e acy and AI implemen a ion abili y.
In e es ingly, he posi i e impac o TMT AI li e acy on AI
implemen a ion abili y is mo e p onounced in s a ups com-
pa ed o incumben i ms. This esea ch en iches he uppe
echelons li e a u e by in oducing AI li e acy as a c i ical
skill-based dimension o TMTs, complemen ing exis ing
ole-o ien ed pe spec i es. I also elucida es he mechanisms
h ough which AI li e acy in op managemen con ibu es o
AI-d i en alue c ea ion wi hin i ms.
The hi d a icle in his opical collec ion, Seeking Empa-
hy o Sugges ing a Solu ion? E ec s o Cha bo Messages
on Se ice Failu e Reco e y, by Ma in Haup , Anna Rozu-
mowski, Jan F eidank, and Alexande Haas (Haup e al.,
2023), in es iga es he use o ailu e eco e y messages in
cha bo s o imp o e use sa is ac ion and e-use in en ions
Elec onic Ma ke s (2024) 34:52 Page 7 o 8 52
ollowing unsuccess ul in e ac ions. As cha bo s a e inc eas-
ingly employed o digi al cus ome in e ac ions, hei e-
quen inabili y o p o ide app op ia e esponses can lead o
use dissa is ac ion, nega i ely impac ing he i m’s se ice
pe o mance. D awing on he s e eo ype con en model, he
au ho s examine he e ec s o wo ypes o ailu e eco e y
messages—solu ion-o ien ed and empa hy-seeking—on
use s’ pos - eco e y sa is ac ion. Th ough h ee expe i-
men s, he s udy inds ha eco e y messages posi i ely
in luence use s’ esponses, media ed by social cogni ions.
Speci ically, solu ion-o ien ed messages enhance pe cep-
ions o compe ence, while empa hy-seeking messages
inc ease pe cep ions o wa m h. The s udy u he e eals
ha he p e e ence o ei he message ype is in luenced by
how use s a ibu e he ailu e and he equency o such
ailu es. These indings o e aluable insigh s o cha bo
de elope s and ma ke e s on how o design e ec i e eco -
e y s a egies ha main ain use sa is ac ion and encou age
con inued use, he eby enhancing cus ome expe ience wi h
digi al con e sa ional agen s in a cos -e ec i e manne . This
esea ch highligh s he impo ance o ailo ed communica-
ion s a egies in mi iga ing he nega i e impac s o cha bo
ailu es and os e ing posi i e use expe iences.
The ou h a icle in his opical collec ion, AI-based
Cha bo s in Con e sa ional Comme ce and Thei E ec s
on P oduc and P ice Pe cep ions, by Jus ina Sidlausk-
iene, Yannick Joye and Vil e Au uske iciene (Sidlauskiene
e al., 2023), explo es he impac o an h opomo phic e bal
design cues in AI-based cha bo s on consume pe cep ions o
p oduc pe sonaliza ion and hei willingness o pay highe
p ices in con e sa ional comme ce con ex s. Al hough
ad ancemen s in na u al language p ocessing (NLP) and AI
a e changing shopping beha io s, many consume s s ill p e-
e human in e ac ions o e cha bo s, which a e o en seen as
impe sonal. The s udy add esses his challenge by examining
how human-like cha ac e is ics in cha bo communica ion
can enhance he shopping expe ience. Th ough a p e- es
and wo online expe imen s, he au ho s ind ha an h opo-
mo phism signi ican ly enhances pe cei ed p oduc pe son-
aliza ion. Addi ionally, his e ec is mode a ed by si ua ional
loneliness, indica ing ha consume s who eel lonely a e
mo e esponsi e o an h opomo phic cues. The in e ac ion
be ween an h opomo phism and si ua ional loneliness also
in luences consume s’ willingness o pay a highe p ice o
p oduc s. These indings sugges ha inco po a ing human-
like e bal elemen s in cha bo design can imp o e consume
engagemen and sa is ac ion, pa icula ly o hose expe i-
encing si ua ional loneliness. The s udy p o ides de elop-
e s and ma ke e s wi h aluable insigh s in o he s a egic
choices when adop ing cha bo s wi h human ai s, bu also
sheds mo e ligh on he psychological dynamics be ween
loneliness and non-human en i ies ha need o be c i ically
e lec ed upon.
We would like o hank he au ho s o hei con ibu-
ions, he e iewe s o hei aluable and p omp eedback,
and Elec onic Ma ke s o making his opical collec ion
possible. These join e o s ha e enabled us o p esen new
esea ch indings in he apidly e ol ing ield o AI-enabled
in o ma ion sys ems.
Funding Open Access unding enabled and o ganized by P ojek
DEAL.
Open Access This a icle is licensed unde a C ea i e Commons A i-
bu ion 4.0 In e na ional License, which pe mi s use, sha ing, adap a-
ion, dis ibu ion and ep oduc ion in any medium o o ma , as long
as you gi e app op ia e c edi o he o iginal au ho (s) and he sou ce,
p o ide a link o he C ea i e Commons licence, and indica e i changes
we e made. The images o o he hi d pa y ma e ial in his a icle a e
included in he a icle’s C ea i e Commons licence, unless indica ed
o he wise in a c edi line o he ma e ial. I ma e ial is no included in
he a icle’s C ea i e Commons licence and you in ended use is no
pe mi ed by s a u o y egula ion o exceeds he pe mi ed use, you will
need o ob ain pe mission di ec ly om he copy igh holde . To iew a
copy o his licence, isi h p://c ea i ecommons.o g/licenses/by/4.0/.
Re e ences
Bailey, D., Fa aj, S., Hinds, P., on K ogh, G., & Leona di, P. (2019).
Special issue o o ganiza ion science. Eme ging echnologies and
o ganizing. O ganiza ion Science, 30(3), 642–646. h ps:// doi. o g/
10. 1287/ o sc. 2019. 1299
Bai d, A., & Ma uping, L. M. (2021). The nex gene a ion o esea ch
on IS use: A heo e ical amewo k o delega ion o and om
agen ic IS a i ac s. Managemen In o ma ion Sys ems Qua -
e ly,45(1), 315–341. h ps:// doi. o g/ 10. 253 00/ misq/ 2021/ 15882
Banh, L., & S obel, G. (2023). Gene a i e a i icial in elli-
gence. Elec onic Ma ke s, 33, 63. h ps:// doi. o g/ 10. 1007/
s12525- 023- 00680-1
Benbya, H., Da enpo , T. H., & Pachidi, S. (2020). A i icial in el-
ligence in o ganiza ions: Cu en s a e and u u e oppo uni ies.
MIS Qua e ly Execu i e, 19(4), 4. h ps:// doi. o g/ 10. 2139/ ss n.
37419 83
Be en e, N., Gu, B., Recke , J., & San hanam, R. (2021). Managing
a i icial in elligence. Managemen In o ma ion Sys ems Qua -
e ly, 45(3), 1433–1450. h ps:// doi. o g/ 10. 25300/ MISQ/ 2021/
16274
Be g, J. M., Raj, M., & Seamans, R. (2023). Cap u ing alue om
a i icial in elligence. Academy o Managemen Disco e ies, 9(4),
424–428. h ps:// doi. o g/ 10. 5465/ amd. 2023. 0106
B on enb enne , U. (1979). The ecology o human de elopmen . Expe -
imen by na u e and design. Ha a d Uni e si y P ess.
Buck, C., Cla ke, J., To es de Oli ei a, R., Desouza, K. C., &
Ma ou khani, P. (2023). Digi al ans o ma ion in asse -in ensi e
o ganisa ions: The ligh and he da k side. Jou nal o Inno a-
ion, Knowledge, 8(2),100335. h ps:// doi. o g/ 10. 1016/j. jik. 2023.
100335
Desouza, K. C., & Dawson, G. S. (2023). Doing s a egic in o ma-
ion sys ems esea ch o public alue. The Jou nal o S a egic
In o ma ion Sys ems, 32(4),101805. h ps:// doi. o g/ 10. 1016/j. jsis.
2023. 101805
Elec onic Ma ke s (2024) 34:5252 Page 8 o 8
Duda, S., Ho mann, P., U bach, N., Völ e , F., & Zwickel, A. (2024).
The impac o esou ce alloca ion on he machine lea ning li e-
cycle. Business & In o ma ion Sys ems Enginee ing,66(2), 203–
219. h ps:// doi. o g/ 10. 1007/ s12599- 023- 00842-7
Eu opean Commission (2019). E hics guidelines o us wo hy AI.
A ailable online a h ps:// da a. eu opa. eu/ doi/ 10. 2759/ 346720
Fue s enau, D., Ro he, H. (2014). Shadow IT sys ems. Disce ning he
good and he e il. ECIS 2014 P oceedings.
Guggenbe ge , T., Lämme mann, L., U bach, N., Wal e , A. M., &
Ho mann, P. (2023) Task delega ion om AI o humans. A p in-
cipal-agen pe spec i e. In P oceedings o he 44 h In e na ional
Con e ence on In o ma ion Sys ems.
Haup , M., Rozumowski, A., F eidank, J., & Haas, A. (2023). Seeking
empa hy o sugges ing a solu ion? E ec s o cha bo messages on
se ice ailu e eco e y. Elec onic Ma ke s,33, 56. h ps:// doi.
o g/ 10. 1007/ s12525- 023- 00673-0
Hinsen, S., Ho mann, P., Jöhnk, J., & U bach, N. (2022). How can
o ganiza ions design pu pose ul human-AI in e ac ions. A p ac-
ical pe spec i e om exis ing use cases and in e iews. In 55 h
Hawaii In e na ional Con e ence on Sys em Sciences.
Jakob, A., Schüll, M., Ho mann, P., & U bach, N. (2024) Teaming up
wi h in elligen agen s. A wo k sys em pe spec i e on he collabo-
a ion wi h in elligen agen s. ECIS 2024 P oceedings.
Janiesch, C., Zschech, P., & Hein ich, K. (2021). Machine lea ning and
deep lea ning.Elec onic Ma ke s,31(3), 685–695. h ps:// doi.
o g/ 10. 1007/ s12525- 021- 00475-2
Jennings, N. R., & Woold idge, M. (Eds.). (1998). Agen echnology.
Sp inge .
Jöhnk, J., Weiße , M., & Wy ki, K. (2021). Ready o no , AI comes.
An in e iew s udy o o ganiza ional AI eadiness ac o s. Busi-
ness and In o ma ion Sys ems Enginee ing,63(1), 5–20. h ps://
doi. o g/ 10. 1007/ s12599- 020- 00676-7
Lämme mann, L., Ho mann, P., & U bach, N. (2024). Managing a i i-
cial in elligence applica ions in heal hca e. P omo ing in o ma ion
p ocessing among s akeholde s. In e na ional Jou nal o In o -
ma ion Managemen ,75, 102728. h ps:// doi. o g/ 10. 1016/j. ijin
omg . 2023. 102728
Leibig, C., B ehme , M., Bunk, S., Byng, D., Pinke , K., & Umu lu, L.
(2022). Combining he s eng hs o adiologis s and AI o b eas
cance sc eening. A e ospec i e analysis. The Lance Digi al
Heal h,4(7), e507–e519. h ps:// doi. o g/ 10. 1016/ s2589- 7500(22)
00070-x
Leona d-Ba on, D. (1995). Wellsp ings o knowledge: Building and
sus aining he sou ces o inno a ion. Ha a d Business Re iew
P ess.
Maddali, H. T., Dixon, E., P adhan, A., & Laza , A. (2022) In es iga -
ing he po en ial o a i icial in elligence powe ed in e aces o
suppo di e en ypes o memo y o people wi h demen ia. In
Ex ended abs ac s on Human ac o s in compu ing sys ems. CHI
Con e ence(pp. 1–7). h ps:// doi. o g/ 10. 1145/ 34911 01. 35198 58
Män ymäki, M., Minkkinen, M., Bi ks ed , T., & Viljanen, M. (2022).
Pu ing AI e hics in o p ac ice: The hou glass model o o ganiza-
ional AI go e nance. a Xi p ep in a Xi :2206.00335. h ps://
doi. o g/ 10. 48550/ a Xi . 2206. 00335
Ma essich, P. W., & Monsey, B. R. (1992). Collabo a ion: Wha makes
i wo k a e iew o esea ch li e a u e on ac o s in luencing suc-
cess ul collabo a ion. Amhe s H. Wilde Founda ion.
McA ee, A. (2024). Gene ally as e : The economic impac o gen-
e a i e AI. A ailable online a h ps:// blog. google/ echn ology/
ai/a- ne w- epo - e xplo es- he- econo mic- impac - o - gene a i e- ai/.
Accessed24.09.2024
Mölle s, M., Be ge , B., & Klein, S. (2024). Con as ing Human-AI
Wo kplace Rela ionship Con igu a ions. In I. Cons an iou, M. P.
Joshi, & M. S elmaszak (Eds.), Resea ch Handbook on A i icial
In elligence and Decision Making in O ganiza ions (pp. 282–303).
Edwa d Elga Publishing.
O likowski, W. J., & Ya es, J. (2002). I ’s abou ime: Tempo al s uc-
u ing in o ganiza ions. O ganiza ion Science,13(6), 684–700.
h ps:// doi. o g/ 10. 1287/ o sc. 13.6. 684. 501
Pinski, M., Ho mann, T., & Benlian, A. (2024). AI Li e acy o he
op managemen : An uppe echelons pe spec i e on co po a e AI
o ien a ion and implemen a ion abili y. Elec onic Ma ke s,34,
24. h ps:// doi. o g/ 10. 1007/ s12525- 024- 00707-1
Sidlauskiene, J., Joye, Y., & Au uske iciene, V. (2023). AI-based cha -
bo s in con e sa ional comme ce and hei e ec s on p oduc and
p ice pe cep ions. Elec onic Ma ke s,33, 24. h ps:// doi. o g/ 10.
1007/ s12525- 023- 00633-8
Te een, L. G. (1995). O e iew o human-compu e collabo a ion.
Knowledge-Based Sys ems,8(2–3), 67–81. h ps:// doi. o g/ 10.
1016/ 0950- 7051(95) 98369-H
Wang, D., Shneide man, B., Chu chill, E., Shi, Y., Maes, P., Fan, X., &
Wang, Q. (2020). F om Human-Human collabo a ion o Human-
AI collabo a ion. Designing AI sys ems ha can wo k oge he
wi h people. In CHI 2020: CHI Con e ence on Human Fac o s in
Compu ing Sys ems (pp. 1–6).
Wei h, H., & Ma , C. (2023). In o ma ion p o ision measu es o
oice agen p oduc ecommenda ions— The e ec o p o-
cess explana ions and p ocess isualiza ions on ai ness pe -
cep ions. Elec onic Ma ke s,33, 57. h ps:// doi. o g/ 10. 1007/
s12525- 023- 00668-x
Wesche, J. S., & Sonde egge , A. (2019). When compu e s ake he
lead: The au oma ion o leade ship. Compu e s in Human Beha -
io ,101, 197–209. h ps:// doi. o g/ 10. 1016/j. chb. 2019. 07. 027
Publishe 's No e Sp inge Na u e emains neu al wi h ega d o
ju isdic ional claims in published maps and ins i u ional a ilia ions.
