LUMI AI Fac o y Se ice Cen e
Empowe ing Eu ope’s AI Ecosys em
D2.2 Cus ome Needs Analysis Templa e
2
D2.2
Cus ome Needs Analysis Templa e
D2.2 Cus ome Needs Analysis Templa e
3
P ojec Ti le
LUMI AI Fac o y Se ice Cen e
P ojec Ac onym
LUMI-AIF
P ojec Numbe
101234208
Type o Ac ion
HORIZON-JU-RIA
Topic
HORIZON-JU-EUROHPC-2025-AI-01-IBA-01
S a ing Da e o P ojec
01.03.2025
Ending Da e o P ojec
29.02.2028
Du a ion o he P ojec
36 mon hs
Websi e
lumi-ai- ac o y.eu
Wo k Package
WP2 Cus ome Engagemen
Task
Task 2.3 Cus ome needs analysis and acili a ing access o AI
supe compu e s
Lead Au ho s
Ou i Tasala (CSC)
Con ibu o s
Susanne Sjöblom (CSC)
Ma in Duda (IT4I)
Ul ike Michel-Schneide (IT4I)
Pe a Lycko a (IT4I)
Pee Re iewe s
Dan S ill (CSC)
Pa ycja Dab iwska-Wie zbicka (Cy one )
Ve sion
1.0
Due Da e
30.6.2025
Submission Da e
27.6.2025
Dissemina ion le el
x
PU: Public
SEN: Sensi i e – limi ed unde he condi ions o he G an Ag eemen
EU-RES. Classi ied In o ma ion: RESTREINT UE (Commission Decision 2005/444/EC)
EU-CON. Classi ied In o ma ion: CONFIDENTIEL UE (Commission Decision 2005/444/EC)
EU-SEC. Classi ied In o ma ion: SECRET UE (Commission Decision 2005/444/EC)
D2.2 Cus ome Needs Analysis Templa e
4
Ve sion His o y
Re ision
Da e
Edi o s
Commen s
0.1
12.6.2025
Ou i Tasala,
Susanne Sjöblom,
Ini ial e sion sen o in e nal e iew
0.2
24.6.2025
Ou i Tasala
I e a ed e sion sen o Se ice
Managemen Boa d o commen s
0.3
26.6.2025
Ou i Tasala
Final e sion sen o PMO o quali y check
1.0
27.6.2025
Anna Luoma
Quali y checked pe o med by he PMO and
sen o o icial e iew.
Glossa y o Te ms
I em
Desc ip ion
Cus ome
An o ganiza ion ha uses LUMI AI Fac o y se ices.
Cus ome
manage
A pe son esponsible o he day- o-day cus ome engagemen and
sales ope a ions. Ac s as he p ima y con ac o assigned cus ome s.
Cus ome needs
su ey
A su ey ca ied ou ei he independen ly by he cus ome ( o m) o
wi h he assis ance o AI Fac o y s a (ea ly discussion) o assess he
needs o he se ices o o m a base o he Cus ome Needs Analysis.
Cus ome needs
analysis
Ac ion ha is ei he au oma ed o ca ied ou by he AI Fac o y s a o
iden i y he bes i o se ices o he cus ome 's need. Analysis is o en
based on he su ey and a discussion ollowing ha .
Digi aliza ion
Digi aliza ion e e s o he p ocess o using digi al echnologies o
change a business model, p o ide new e enue and alue-p oducing
oppo uni ies, and ans o m how o ganiza ions ope a e and deli e
alue o hei cus ome s. Digi aliza ion in ol es in eg a ing digi al
echnologies in o e e yday li e and business p ocesses.
Se ice ca alogue
Cus ome - acing lis o all li e se ices o e ed along wi h ele an
in o ma ion abou hese se ices.
Se ice po olio
In e nal lis ha de ails all he se ices o e ed by a se ice p o ide ,
including hose in p epa a ion, li e and discon inued.
Use
Pe son wi h a use accoun in LUMI o LUMI-AI, ob ained ei he om
he cus ome po al o ia he Eu oHPC Fede a ion Pla o m.
D2.2 Cus ome Needs Analysis Templa e
5
Execu i e Summa y
LUMI AI Fac o y suppo s he cus ome jou ney by conduc ing s uc u ed cus ome needs analyses o
align he se ices o e ing wi h he speci ic equi emen s and eadiness le els o each o ganiza ion.
These analyses a e guided by a s anda dized ool designed o assess he cus ome ’s cu en expe ise
and capabili ies. This enables ailo ed ecommenda ions, such as whe he o begin wi h ounda ional
HPC aining o p oceed di ec ly o applying o compu ing and da a esou ces o launch an AI p ojec .
This documen ou lines he de ini ion and pu pose o he cus ome needs analysis, de ailing i s ole in
enhancing cus ome engagemen . I also p esen s he me hodology used and in oduces he su ey
empla e employed in he assessmen p ocess. Fu he mo e, i ou lines he planned de elopmen
ac i i ies aimed a e ining he cus ome needs analysis h oughou he cou se o he p ojec .
The cus ome needs analysis is buil on a cus ome - i s app oach, ensu ing ha each o ganiza ion is
guided owa d he mos app op ia e se ices. Insigh s ga he ed om ea ly su ey esponses will in o m
con inuous imp o emen o he p ocess, ensu ing i emains esponsi e and e ec i e.
D2.2 Cus ome Needs Analysis Templa e
6
Table o Con en s
1. Objec i e .................................................................................................................... 7
2. Scope ......................................................................................................................... 8
3. Bene i s o he cus ome ......................................................................................... 10
4. Me hodology o mapping Cus ome Needs .............................................................. 11
5. Cus ome Needs Su ey empla e s uc u e and classi ica ion amewo k ................. 12
6. Cus ome Needs Analysis e iewed and de eloped ................................................... 18
7. Conclusions .............................................................................................................. 19
D2.2 Cus ome Needs Analysis Templa e
7
1. Objec i e
The objec i e o he Cus ome Needs Analysis (CNA) is o unde s and he speci ic equi emen s, pain
poin s, and goals o he cus ome s who a e le e aging o planning o le e age AI and HPC echnologies.
Ra he han being a s andalone se ice, he CNA se es as a s a egic ool o assess a cus ome 's skills,
and echnical eadiness. The analysis sheds ligh on he e o s and asse s needed o le e age AI and
HPC echnologies. I also desc ibes he company’s eadiness, skills and capabili ies in AI and HPC
de elopmen and shows he skills gaps and compe ence needs and gi es a oad map o ac i i ies
needed in he RDI de elopmen pa h. I helps ma ch use s wi h se ices bes sui ed o hei si ua ion,
whe he hey a e new o AI o HPC o eady o scale up p ojec s. Fo ins ance, he cus ome needs
analysis can de e mine i a cus ome should begin wi h in oduc o y HPC aining o p oceed di ec ly o
applying o compu e and da a esou ces o un an AI-d i en wo kload. Thus, he main pu pose o he
cus ome needs analysis is o help he cus ome na iga e wi hin he ex ensi e se ice po olio o he
LUMI AI Fac o y.
The cus ome needs analysis is an in eg al pa o he cus ome p ocess in LUMI AI Fac o y. Cus ome
jou ney and he ela ed se ice o e ing om he LUMI AI Fac o y is desc ibed in mo e de ail in
deli e able 2.1 Cus ome P ocess Desc ip ion and hus only he simpli ied g aph o he cus ome jou ney
is p esen ed he e o e e ence:
Figu e 1 Cus ome Jou ney
D2.2 Cus ome Needs Analysis Templa e
8
The cus ome needs analysis aims o:
• Align se ice o e ing wi h cus ome needs and goals
• Guide cus ome s owa d app op ia e aining, ools, and esou ces
• In o m he de elopmen and e inemen o LUMI AI Fac o y se ices
The su ey ocuses on cus ome s’ needs and expec a ions and hei ba ie s and eadiness in using AI
Fac o y se ices, also ouching upon hei ma u i y in HPC and AI. In he cu en cus ome p ocess he
LUMI AI Fac o y se ice ca alogue is g ouped in o i e main ca ego ies, each designed o mee he
speci ic needs and ma u i y le els o use s. These ca ego ies a e s ill subjec o change bu a he ime
o w i ing his deli e able a e: Onboa ding & Sel -se ice Resou ces, Technical & Sys em Suppo ,
Pe sonalized se ices (Expe Consul ing & P ojec Suppo ), Skills De elopmen and T aining, Access
o Da a and Co-wo king Spaces & Collabo a ion Se ices, in oduced in de ail in deli e able 2.1
Cus ome P ocess Desc ip ion. The se ice ca alogue will be accessible ia he websi e www.lumi-ai-
ac o y.eu and in he u u e h ough a po al. The se ice o e ing is in oduced o he cus ome s
h ough Cus ome Needs Analysis Su ey. The su ey p oduces an au oma ed ecommenda ion i s on
he sugges ed se ice ca ego ies and la e indi idual se ices om he ca ego ies. These au oma ed
ecommenda ions can be ollowed up wi h an in-dep h discussion o p oduce he mos alue o he
cus ome .
Wi h he aid o he cus ome needs analysis, we can agg ega e and analyze da a ac oss cus ome s o
iden i y la ge ends and in u n be e help he u u e cus ome s. We can also gene a e epo s on
in e nal de elopmen o s akeholde communica ion based on he insigh s.
Al hough he ocus is o iden i y and mee he needs o he cus ome wi h he LUMI AI Fac o y se ices,
cus ome needs analysis s i es o se e he cus ome also in he case LUMI AI Fac o y does no di ec ly
p o ide he se ice equi ed. As pa o he ope a ion, we aim o ma ch he cus ome wi h di e en
ac o s and ecosys ems, aiding hem along in hei AI jou ney ega dless.
As pa o he ini ial ollou , he i s 50 Cus ome Needs Analyses will be comple ed by he end o M10,
and in doing so u he insigh will be gained in o how o de elop he posi ioning o he cus ome cases
and o be e ma ch he se ices wi h he cus ome s. The cus ome needs analysis empla e is also
a ailable o be used o he deli e able 7.1 Su ey o cus ome needs and HPC/AI ma u i y.
2. Scope
Cus ome Needs Analysis is a ool and se ice o he cus ome o e ed in s ep 2c in he Cus ome
Jou ney g aph, shown in chap e 1 o his deli e able and in oduced in de ail in chap e 5 o he
deli e able 2.1 Cus ome P ocess Desc ip ion. Cus ome needs analysis will be ca ied ou i s and
o emos o hose cus ome s who con ac o o he wise need he help o LUMI AI Fac o y in inding
sui able se ices o i hei needs. Thus, i a cus ome equi es access o speci ic o ailo ed se ices,
hey will always be o e ed cus ome needs analysis. The cus ome needs analysis will no be
p oac i ely o e ed o he cus ome s who u ilize he eely o e ed onboa ding and sel -se ices (s ep
D2.2 Cus ome Needs Analysis Templa e
9
2a) no he ones applying o HPC esou ces independen ly ei he h ough Eu oHPC JU access calls o
he na ional access models (s ep 2b). Howe e , he o m is eadily accessible a he onboa ding s age
also o cus ome s h ough s eps 2a and 2b.
Cus ome needs analysis ool package includes:
• su ey o e ed ia a o m, as he ini ial s ep whe e he cus ome p o ides inpu in a s uc u ed
o m
• au oma ed o guided ecommenda ion o se ices based on he ini ial sco ing o he su ey
• in-dep h discussion ollowing he su ey empla e o discuss he cus ome need u he , as
needed
The cus ome needs su ey is accessible om he on page o he LUMI AI Fac o y websi e www.lumi-
ai- ac o y.eu and in he u u e h ough a po al, bu he cus ome is also di ec ed o i om he se ice
ca alogue, o ganized by cus ome needs. The su ey is in oduced ia a o m on he websi e and he
cus ome ills i in when lea ing a se ice eques . The su ey is planned o be sho enough o highligh
some main aspec s o he use , s ill p o iding enough insigh in o he cus ome need o enable he AI
Fac o y o espond o he cus ome need in he bes possible way. LUMI AI Fac o y se ice po olio is
p ominen ly showcased also a indus y e en s. When possible, on- he-spo cus ome need
assessmen s a e o e ed, lowe ing he h eshold o new cus ome s o connec wi h ou se ices and
helping hem iden i y conc e e oppo uni ies o collabo a ion. The su ey is a ool bu also a se ice
p o ided o he use o align hei needs wi h he se ices.
Based on he su ey, an au oma ed e alua ion is conduc ed o de e mine whe he he cus ome ’s
needs can be me by he LUMI AI Fac o y se ices. This assessmen conside s business, echnical, and
egula o y pe spec i es. Cus ome ecei es ins an ecommenda ions o he AI Fac o y se ices ha
migh bes mee he needs o ha cus ome . In cases whe e he LUMI AI Fac o y is no he bes i , he
cus ome may be e e ed o a mo e app op ia e se ice p o ide o pa ne .
The ou come o he Cus ome needs analysis guides he use owa d he app op ia e en y poin and
se ice use — whe he sel -se ice esou ces, aining, o p ojec -based collabo a ion. Ce ain se ices,
such as mos aining cou ses, e en s, and access o co-wo king spaces, may be eely a ailable o all
cus ome s om Eu oHPC membe s a es. In case he cus ome needs analysis indica es ha he need is
o ailo ed o ad anced HPC and AI esou ces, inding he sui able pa hways will be u he suppo ed
by an op ional HPC and AI Ma u i y e alua ion. Ge ing access o ex ensi e compu e esou ces, he
cus ome is equi ed o demons a e he abili y o ei he possess o de elop he necessa y skills o
success ully execu e HPC o AI p ojec s. HPC and AI Ma u i y e alua ion assesses on a deepe le el he
cus ome 's echnical eadiness and AI capabili ies, ensu ing ha se ices a e well aligned wi h he
cus ome ’s de elopmen s age. This ma u i y e alua ion is closely ied and a supplemen o he
cus ome needs analysis. The con en and s uc u e o he HPC and AI Ma u i y Su ey a e desc ibed in
de ail in deli e able 2.3 HPC and AI Ma u i y Su ey.
A e he Cus ome Needs Analysis (and possible HPC and AI ma u i y e alua ion) cus ome needs-
d i en ma ching is suppo ed by LUMI AI Fac o y se ice expe s, p o iding business domain expe ise,
as well as pe sonalized cus ome suppo by he Cus ome Manage s. Ca ego ies o se ices a ailable
a e he cus ome needs analysis include Pe sonalized se ice (Expe Consul ing & P ojec Suppo ),
Skills De elopmen & T aining, Access o Da a, and Co-wo king spaces and Collabo a ion Se ices. A
D2.2 Cus ome Needs Analysis Templa e
16
• Basic AI/HPC aining (e.g., in oduc ion o machine and deep lea ning)
• Ad anced AI/HPC aining (e.g., scaling and op imizing AI jobs)
• Domain-speci ic aining (e.g., unning domain-speci ic so wa e o asks on LUMI)
• Online cou ses
• In-pe son wo kshops
• Indi idual consul a ions
• No in e es
Se ice ecommenda ions a e made om he Skills de elopmen and T aining ca ego y acco ding o
he esponses.
A e you in e es ed in using co-wo king spaces o s uden alen p og ams?
• Yes, main hub (Aal o Uni e si y)
• Yes, sa elli e hubs
• No
Cus ome is poin ed o he closes AI Hub, i applicable.
In which a eas would you welcome consul a ion suppo ? (Selec all ha apply)
• Feasibili y analysis o AI p ojec s
• Applying o la ge compu a ional esou ces
• Regula o y/compliance/e hics (e.g. us wo hy AI, GDPR)
• Ecosys em and pa ne ship de elopmen
• O he (please speci y)
• None
Se ice ecommenda ions a e made om he Pe sonalized se ice ca ego y acco ding o he
esponses.
D) Ba ie s and Readiness
Pu pose: To iden i y obs acles and eadiness o AI adop ion.
Wha a e you main ba ie s o AI adop ion? (Rank up o h ee)
• Lack o expe ise
• Lack o da a
• Budge cons ain s
• Regula o y unce ain y
• Insu icien in as uc u e
• Unclea business alue
• O he (please speci y)
• None
D2.2 Cus ome Needs Analysis Templa e
17
Se ice ecommenda ions a e made om he Onboa ding & Sel -se ice, Skills de elopmen and
T aining, Access o Da a, Pe sonalized se ice and Technical & Sys em Suppo ca ego ies acco ding o
he esponses.
How do you assess you company’s eadiness o adop and use AI?
Response
Sco e
Readiness le el
No eady
1
Low
S ill explo ing possibili ies
2
Eme ging
Pilo p ojec s, bu no o e all
s a egy
3
Medium
Clea s a egy and leade ship
4
High
Fo answe s wi h a sco e o 2 o highe , a sugges ion o comple ing he HPC and AI ma u i y e alua ion
is made.
E) Mo i a ion and Ou look
Pu pose: To cla i y mo i a ion and expec ed ou comes.
Wha a e you main mo i a ions o engaging wi h AI Fac o y? (Selec all ha apply + open ques ion)
• Reduce ime o ma ke
• De elop new p oduc s o se ices
• Inc ease e iciency
• Launch new p oduc s/se ices
• De elop exis ing p oduc s/se ices
• Reduce cos s
• Imp o e cus ome expe ience
• Mee egula o y equi emen s
• O he (please speci y)
• I don' know
The cus ome will be asked o gi e a sho desc ip ion based on he answe .
Wha bene i s a e you looking om he use AI Fac o y se ices?
• Inc eased speed o da a p ocessing
• Simula ion o po en ial mal unc ion, bo lenecks o occu ence
• Simula ion o pe o mance wi h a la ge amoun o pa ame e s/s o ed da a
• Big da a analysis
• Applica ion o hea y ma hema ical algo i hms o big da a
• Moni o ing and imp o ing en i onmen al impac
D2.2 Cus ome Needs Analysis Templa e
18
• P edic i e models
• De eloping AI-models
• Some hing else
• I don' know
Se ice ecommenda ions a e made om he Onboa ding & Sel -se ice, Skills de elopmen and
T aining, Access o Da a, Pe sonalized se ice and Technical & Sys em Suppo ca ego ies acco ding o
he esponses.
Wha is you ime ho izon o you i s (o nex ) AI p ojec ?
• Immedia ely
• Wi hin 6 mon hs
• Wi hin 1 yea
• No conc e e plans ye
A e you in e es ed in long- e m collabo a ion o one-o suppo ?
• Long- e m pa ne ship
• One-o suppo
• No su e
F) Open Feedback (Op ional)
Pu pose: To cap u e quali a i e insigh s o special needs no co e ed abo e.
Please desc ibe any speci ic expec a ions o needs you ha e ega ding AI Fac o y se ices:
• Open ex
Open eedback is ga he ed as an inpu on he Cus ome Needs Analysis i sel and o aid he
de elopmen o he se ices. Responses can lead o in-dep h discussions wi h a Cus ome Manage o
explo e u he he needs and possibili ies p o ided by he LUMI AI Fac o y.
6. Cus ome Needs Analysis e iewed and de eloped
The su ey will be es ed on an ini ial se o cus ome s. Based on hese e y i s indings, i e a i e
upda es o he Cus ome Needs Analysis a e planned. A e he ini ial phase, he empla e will be
e iewed and, i deemed necessa y, adjus ed yea ly, based on he expe ience gained om he analysis.
Cus ome Needs Analysis me hods and ools will also be e alua ed in he ligh o he o e all eedback
on he onboa ding phase and AI Fac o y se ices in suppo ing he cus ome .
Fi s he ocus will be on suppo ing he cus ome a he ini ial con ac and onboa ding phase, bu he
a enues o suppo he cus ome s mo e ho oughly ia he ools o he cus ome needs analysis a he
D2.2 Cus ome Needs Analysis Templa e
19
la e s ages o he cus ome jou ney will be in es iga ed u he . How he me hod could be be e
u ilized o ack he e ol ing cus ome needs will be assessed. T acking cus ome p og ess and ac ual
se ice use and assessing hese agains he ini ial au oma ed ecommenda ions could be examined in
he u u e. Di e en cus ome p o iles could be c ea ed based on he segmen a ion in o ma ion, he
needs, capabili ies and ba ie s iden i ied om he su ey da a.
7. Conclusions
The Cus ome Needs Analysis se es as a co ne s one in aligning he LUMI AI Fac o y’s se ice o e ings
wi h he e ol ing needs, eadiness, and ambi ions o i s di e se cus ome base. By sys ema ically
iden i ying cus ome needs, ba ie s o he use o se ices and cus ome ’s capabili ies, Cus ome
Needs Analysis enables a ailo ed app oach o se ice deli e y—ensu ing ha each cus ome is guided
owa d he mos app op ia e ools, aining, and esou ces.
Cus ome Needs Analysis is a s a egic ins umen ha suppo s con inuous de elopmen . I no only
acili a es pe sonalized ecommenda ions bu also con ibu es o he b oade e inemen o he AI
Fac o y’s se ices by e ealing ends, skill gaps, and eme ging needs ac oss he ecosys em. This dual
ole—suppo ing indi idual cus ome s while in o ming se ice e olu ion—makes he Cus ome Needs
Analysis a i al componen o he LUMI AI Fac o y cus ome p ocess, as de ailed in deli e able 2.1.
Mo eo e , he Cus ome Needs Analysis’ abili y o connec cus ome s wi h ex e nal ac o s and
ecosys ems, e en when hei needs ex end beyond he cu en scope o LUMI AI Fac o y, unde sco es
i s alue as a holis ic suppo mechanism. As he ini ial ollou p og esses and insigh s om he i s 50
analyses a e ga he ed, he Cus ome Needs Analysis will con inue o e ol e, enhancing i s abili y o
posi ion cus ome cases e ec i ely and op imize se ice ma ching.
Ul ima ely, he Cus ome Needs Analysis empowe s cus ome s o na iga e he complex landscape o
AI and HPC wi h con idence, ensu ing ha hei jou ney— om explo a ion o implemen a ion—is bo h
s a egic and suppo ed.
