AI in Greek Retail
Which business problems are Greek retail leaders trying to solve with AI?
A joint study of retail and commerce leaders: what they want AI to solve, how far they've really got, and what holds them back.
- Edition 01
- In partnership with SELPE
- Fieldwork 9 November to 18 December 2026
- Delphi Economic Forum, April 2027
- M.O.R.E Summit, July 2027
Greek retail talks a lot about AI. Very little of it does real work yet.
BORRN and SELPE create the first study of how Greek retail and commerce uses AI: which business problems leaders want it to solve, how far they have really got, and what holds them back. We'd love to work with SELPE on this initiative: we build it together and present the results together.
What we'll find out
- 01The business problems leaders want AI to solve, ranked.
- 02Where AI already does real work: stores, supply chain, e-commerce, marketing, finance, HR.
- 03How far adoption has gone, from talk to measured results.
- 04What holds companies back, and what leaders do differently.
A research report on the business problems retail leaders want AI to solve, how mature adoption is, and what holds them back.
The first evidence on AI in Greek retail, so the sector can speak about AI with its own numbers.
Owners, CEOs and senior managers of SELPE member companies. Retail and commerce only.
An online survey of 100+ leaders (about 14 minutes) and 8 to 12 interviews of 45 minutes, with answers cross-checked against publicly available information.
Why this study is needed now
- of Greek companies used AI in 2025, against 20% in the EU. Eurostat
- 8.9%
- Greece's rank among the 27 EU countries. Eurostat
- 24th
- of 62 Greek CEOs saw both higher revenue and lower costs from AI. PwC, 2026
- 10%
None of these numbers covers retail. This study will be the first.
- Greece, all firms8.9%
- EU, all firms20.0%
- Greece, large firms37.6%
- EU, large firms55.0%
Who we survey
Owners and CEOs of Greek retail and commerce
- Reached via
- SELPE, which invites its members under its own name
- Size
- At least 100 answers, ideally 150 to 200
- Gives us
- The problems retail leaders want AI to solve, a map of where AI already works by department, and the authority of the sector's own body
A study we build together.
We'd love to work with SELPE on this initiative. We create the study together and present the results together. Executives answer because someone they trust asks them. In research on top managers, an invitation through a known body raised response more than advance notice, follow ups or a personal touch. SELPE brings that trust. BORRN brings the research, the analysis and the report.
Who does what
Builds and runs the research
- Creates the framework, drafts the questionnaire and the communication assets
- Designs, runs and analyses the research, and writes the report
- Owns the raw data
- Prepares private results for members
- Co-creates the launch plan
- Shares the report on BORRN's digital channels
- Co-presents the findings with SELPE at the Delphi Economic Forum
Opens the doors and co-presents
- Invites its members and follows up, with two reminders, until we reach the minimum number of answers for credible research
- Suggests and coordinates the 10 companies for the in-depth interviews
- Reviews the questionnaire and suggests changes
- Co-hosts an event for SELPE members, co-presenting the results
- Shares the report with members, on its channels and with partner media
- Works with BORRN on the M.O.R.E Summit presentation
What SELPE gets
No Greek study today shows how retailers use AI or what they want it to solve. SELPE presents the first numbers on its own sector.
SELPE's logo is on the cover and it is named in every article that covers the report.
The government has announced a €150M scheme for AI aimed at smaller businesses, and the EU is pressing Greece to raise business use of AI from 8.9%. SELPE gets the evidence to speak for retail in that conversation.
Every member that takes part gets private results against its peers and an AI audit from BORRN, and all members get a first look at the findings.
SELPE co-presents the findings at the Delphi Economic Forum in April and at the M.O.R.E Summit in July 2027.
BORRN designs, runs, analyses and writes. SELPE invites, reviews and co-presents.
What we agree in writing
BORRN holds the raw answers. Results are shared as totals only, never for a group of fewer than 10 companies.
SELPE reviews the questionnaire before fieldwork and previews the report before launch. The findings stay independent.
Logos, the cover, and who presents at each event.
Numbers and stories, checked against what we can see.
We ask for both quantitative and qualitative answers. Opinion alone is weak evidence in this market: in July 2026, 75% of Greek CEOs told PwC they were ready for AI, yet only 10% had achieved both revenue growth and cost reduction from it. So we check what leaders tell us against public evidence and audited accounts.
- A lot of opinions about AI
- We collect evidence in four ways
- We turn it into clear findings
- We publish only what is defensible
What leaders say is actually running. SELPE member companies, 24 questions, about 14 minutes online.
Why it is or is not working. 8 to 12 conversations of 45 minutes give the stories behind the numbers.
What the world can already see: annual reports, press, job ads, talks.
Growth and margins from audited accounts, so we can test whether AI leaders also perform better.
- Findings per question
- Headline findings, agreed with SELPE
- The report
One core question, three supporting ones, and our guesses written down before we ask.
A first draft for SELPE to review. Every question is about retail and commerce.
Which business problems are retail leaders trying to solve with AI?
How far has AI adoption in retail really gone, and how mature is it?
What holds retailers back from solving those problems with AI?
Where do retail leaders already see measured results, and what did they do differently?
Hypotheses
Our best guess about the answer, written down before we collect any data, so the data can prove us wrong. Each guess points to the question that tests it. After fieldwork we publish which ones held and which did not. That is what separates research from a brochure.
Retailers aim AI mostly at cost and repetitive work, rarely at growth.
The problem leaders rank first is rarely the one where AI already runs.
Few companies have decided how they will know a problem is solved.
No clear business case
Fragmented data and knowledge
Limited internal capability and ownership
Unclear governance, security and compliance
Tools bought, but the work stays the same
Pilots that never reach measurable value
Companies that brought in outside AI consultants are no further ahead than those that did it with their own people.
Tested by Q19. If it holds, we publish it, even though BORRN is an AI consultancy. Readers trust research that is willing to argue against its own author.
SELPE's members include family retailers and multinationals. Comparing the two could be a headline on its own: are family retailers ahead or behind?
The questionnaire, from broad to specific
24 questions, about 14 minutes. Mostly closed questions for the numbers, with open questions at the end for quotes. We start with who is answering, then the big picture, then narrow into the core question.
AProfile2 min · 6 questions
Who is answering and for which company. It also screens out the wrong respondents.
What is your role?
- Owner
- CEO or MD
- Other C level
- Business unit or functional head
- Other (ends survey)
Which part of retail and commerce is your company in?
- Supermarkets and food
- Fashion and footwear
- Electronics and appliances
- Home, furniture and DIY
- Health, beauty and pharmacy
- Department stores
- Online-only retail
- Wholesale and distribution
- Other retail
How many people does your company employ in Greece?
- Under 50
- 50 to 249
- 250 to 999
- 1,000 or more
What was your annual revenue last year?
- Under €10M
- €10M to €50M
- €50M to €100M
- €100M to €300M
- Over €300M
Who controls the company?
- Its founder
- A family
- Greek owners, not a family
- A multinational group
- Listed, widely held
Which generation leads the company today?
- First
- Second
- Third or later
BThe big picture2 min · 3 questions
Where AI sits in the company overall, asked before any detail so the specific questions do not colour the answer.
How important is AI to your company's plans for the next two years?
- Not on the agenda
- Discussed, no plan
- In the plan, no budget
- In the plan, with budget
- Central to our strategy
Which best describes AI in your company today?
- Not used
- Individuals use tools on their own
- Pilots running
- At least one use runs daily, with a named owner
- Several uses, with measured results
Who leads AI in your company?
- Owner or CEO
- A named executive
- IT
- Each department on its own
- Nobody yet
CThe core: business problems4 min · 5 questions
The heart of the report. We ask about problems first and AI second, so leaders talk about their business, not about technology.
Which business problems are you trying to solve with AI?
- Too much repetitive work
- Stock and demand forecasting
- Pricing and promotions
- Slow decisions or poor information
- Growing sales
- Customer service
- Hiring and keeping store staff
- Compliance and reporting load
- Knowledge held by too few people
- Errors and quality
- Other, write in
Of those, which one matters most?
- Shows only the answers given in Q10
In which department does that problem mainly sit?
- Stores and operations
- Supply chain and logistics
- E-commerce
- Sales and customer
- Marketing
- Finance
- HR
- Legal
- IT
How far have you got with that problem?
- Not started
- Exploring options
- Pilot running
- In daily use, with an owner
- Solved, with a measured result
How will you know it is solved?
- Time saved
- Cost reduced
- Revenue gained
- Fewer errors
- Faster response
- We have not decided
DDepartment by department2 min · 1 question
One grid that becomes the report's signature chart: a map of where AI does real work in Greek retail.
For each department, which best describes AI today? Rows: Stores and operations, Supply chain, E-commerce, Sales and customer, Marketing, Finance, HR, Legal.
- Not used
- Individuals
- Pilot
- Daily, with an owner
- Several, with a KPI
EWhat stands in the way2 min · 4 questions
The barriers, and the facts that test them.
What is the single biggest barrier to solving that problem with AI?
- No clear business case
- Data scattered or poor
- Lack of skills
- Nobody owns it
- Risk, security, compliance
- Tools bought, work unchanged
- Pilots never scale
- Budget
Of the AI licences you pay for, how many are used every week?
- Most
- About half
- Few
- We do not know
- We have none
How will your AI budget for 2027 compare with 2026?
- Lower
- The same
- Higher
- Much higher
- Not decided
Who has helped you put AI to work?
- Our own team
- A software vendor
- A consultancy
- A university
- Nobody yet
FResults1 min · 2 questions
Whether anything has paid off, and how it was measured.
Has any AI use produced a result you measured?
- Yes, reported to the board
- Yes, measured, not reported
- We believe so, not measured
- No
What kind of result?
- Time
- Cost
- Revenue
- Quality
- Customer experience
GClosing1 min · 3 questions
Two open questions for quotes, then consent.
What is the one part of your work you would never hand to AI?
Three years from now, what do you want your company to be known for because of AI?
May we:
- Send you private results against your peers
- Invite you to a 45 minute interview
- Name your company if it is a leader
The numbers tell us what is happening and how much.
The survey sits at the centre, checked by public evidence and audited accounts.
What the survey measures
Ten areas, grouped into five pillars.
- OwnershipOwnership and decision making
- Redesigned workUse cases in production or experiment
- CapabilityData and infrastructure readinessSkills and organisational readiness
- ResultsResults achieved so far
- Scale and controlAdoption and maturityGovernance, risk and trust
- ContextBusiness problems prioritisedInvestment intentionsBarriers to scaling
An option to agree together: a maturity tier for each company
An invented example. Each pillar scores 0 to 100 and counts 20%. The average places the company on one of four shelves. Tiers give members a clear private result and the report clear headlines.
- Scaling75 to 100A system that scales across the company
- Expanding50 to 74Several owned workflows
- Proving25 to 49One real use case
- Starting0 to 24Talk, licences, interest
Four shelves, not a league table: no company #17 against company #43.
How many answers we need
Owners, CEOs and senior managers of retail and commerce companies. Ideally 150 to 200.
With 100 answers a figure is accurate to within about 10 points either way, with 150 about 8, with 200 about 7. For smaller groups, such as one retail category, we report counts instead of percentages.
Reporting rules
Write "9 of 41 companies", never "22%", on any base under 50.
Never show a group of fewer than 10 companies, so no one can be identified.
Compare respondents with SELPE's full membership on category and size, and say so in the report.
"Leaders tend to grow faster," never "AI makes them grow faster."
The findings describe Greek retail and commerce, and say so on every headline.
Test the questionnaire with 5 to 8 people, in Greek and English, and keep it under 15 minutes.
The interviews tell us why, and give the report its voices.
8 to 12 owners or CEOs of SELPE members, across retail categories and maturity levels. Leaders and companies still stuck, so we do not only tell success stories.
45 minutes, in person or by video, in Greek. Recorded only with consent, then transcribed.
Group the answers by theme to explain the survey numbers. The best stories become short case examples in the report.
The conversation, in order
- 1
The business problem you most want AI to solve, and why that one
- 2
The first real AI project and what it changed
- 3
What failed and why
- 4
Who owns AI inside the company
- 5
How results are measured, if at all
- 6
The human line: what part of the work will you never delegate to AI?
- 7
Three years out: what is your team famous for?
- 8
Advice for another retailer starting now
Rules
A quote appears under a name only with written approval of the exact wording.
Everything else stays anonymous.
An interview never turns into a sales pitch.
Every chapter title states a finding as a full sentence.
"AI is doing real work in supply chain at 9 of 41 retailers" tells the reader something. "AI in supply chain" does not. About 28 pages, in Greek and English.
The reader's journey
- 1
Summary. Three to five headline findings on two pages. Most readers stop here.
- 2
Greek retail in context. EU data next to market claims, and why they disagree.
- 3
How far adoption has gone. From talk to measured results.
- 4
The problems retailers want AI to solve, department by department. The core finding, ranked, with a map of where AI already does real work.
- 5
What holds retailers back. Data, people, budget or the owner's attention.
- 6
What leaders do differently. Five or six practices, each with a number.
- 7
What a retailer can do on Monday. One page of concrete steps.
- 8
How we did it. Sample, dates, sources and limits, written for a sceptical reader.
The spotlight rule
The findings map the whole sector.
No public league table, and no member named as a laggard.
Three quotable lines, designed before fieldwork
"X of Y retail owners call AI a priority; Z have it doing measured work."
A name for the leading group that Greek retail can use for years, as Microsoft uses "Frontier Firm".
How leaders differ in growth or margin, from audited accounts.
Outputs and launch
- Full research report, in Greek and English
- Executive summary for senior leaders
- Press release and media briefing
- Presentations for each launch stage
- Social content from the strongest findings, for SELPE and BORRN channels
- Cuts by retail category where the sample allows
- Private results for each participating member
- Soft launch, 27 January 2027
- Event for SELPE members, 1 to 10 February, co-hosted, results co-presented
- Delphi Economic Forum, late April 2027, findings co-presented
- Build on the results, April to May, with deeper cuts for members
- Report launch at the M.O.R.E Summit, July 2027
- Full report free, shared by SELPE with members and by both partners on their channels and partner media
Fieldwork before the holidays, first results in January, the big stages in spring and summer.
The survey and interviews run from 9 November to 18 December. The report is reviewed with SELPE and soft-launched at the end of January, presented first to SELPE members in February, then at the Delphi Economic Forum in April and the M.O.R.E Summit in July. If SELPE confirms later than 30 October, the dates after it move by the same amount.
- OctoberAlignment with SELPE, SELPE confirms
- 9 Nov to 18 DecSurvey and interviews
- JanuaryAnalysis, review with SELPE, soft launch on 27 Jan
- FebruaryFirst presentation to SELPE members
- Late April 2027Delphi Economic Forum
- July 2027Public launch at the M.O.R.E Summit
Suggested timeline. We agree the details together.
Step by step
- Alignment with SELPE on scope, roles and dates 20 to 30 Oct
- SELPE confirms 30 Oct
- Build the questionnaire; SELPE reviews and suggests changes by 6 Nov
- Test with 5 to 8 people and build the online survey by 6 Nov
- Get in touch with the Delphi Economic Forum November to December
- SELPE sends the invitation; the survey goes live 9 Nov
- Two reminders through SELPE, in weeks 2 and 4
- Interviews with 8 to 12 leaders 16 Nov to 18 Dec
- Public evidence collected 2 Nov to 11 Dec
- Survey closes 11 Dec
- Analyse results 7 to 15 Jan
- Write and design the report, in Greek and English 12 to 22 Jan
- Review and confirm the report with SELPE 22 to 27 Jan
- Soft launch 27 Jan
- Launch event for SELPE members 1 to 10 Feb
- Delphi Economic Forum preparation 10 to 28 Feb
- Delphi Economic Forum launch late April
- Build on the results April to May
- Report launch at the M.O.R.E Summit July 2027
- Number and quality of respondents
- C level participation
- SELPE member participation and reach
- Attendance at the launch events
- Downloads of the report
- Relevant media coverage
How we keep the study credible.
In July 2025 an MIT report said 95% of organisations get zero return from generative AI, and headlines around the world repeated it. Then critics read the method: 52 interviews and 153 survey answers collected at four conferences. A report on AI will be read with the same suspicion, so we are open about how we did it.
- 1
Clear authorshipWe say plainly who wrote the report, and publish the full questionnaire, dates and response counts.
Rule - 2
A second pair of eyesSELPE reviews the questionnaire before fieldwork, and we ask an academic for comments on the method.
Rule - 3
Rules written firstWhat counts as public evidence, and how any tier is scored, is written down before we look at the data.
Rule - 4
Consent before namingNo company is named without written consent. All answers are confidential.
Rule - 5
One page on how we did itSample, dates, sources, limits, and how AI was used in the analysis.
Rule - 6
One sector, statedOnly SELPE members answer, so the findings speak for retail and commerce. We say so on the first page and in every headline.
Rule
What we agree together, and what happens next.
To agree together
- 1
Which members we inviteAll members, or the owners and CEOs of a defined group
Open - 2
Family retailers and multinationalsCompare the two, or report them as one group
Open - 3
How many answers we aim forAt least 100, ideally 150 to 200, and the mix of company sizes
Open - 4
Maturity tiersWhether each participating company gets a tier in its private results
Open - 5
The study's name"AI Readiness Index" is already used by Cisco and by ELTRUN with Entersoftone, so we choose a name of our own together
Open - 6
Launch dates and venuesThe soft launch, the members' event, Delphi and M.O.R.E
Open
Next steps
- 1
Alignment with SELPE on scope, roles and the points above
- 2
SELPE confirms by 30 October
- 3
Build the questionnaire and share it with SELPE for review
- 4
Get in touch with the Delphi Economic Forum about co-presenting the findings
Every figure on this page, and where it comes from.
Checked on 16 October 2026 against the original publisher where it was available.
| Figure | Source | Note |
|---|---|---|
| 8.9% of Greek enterprises used AI in 2025; EU 20.0%; large firms 37.6% against 55.0%; Greece 24th of 27 | Eurostat, isoc_eb_ai | Enterprises with 10 or more staff |
| 47% of Greek businesses say they use AI | AWS and Strand Partners, 2026, via To Vima | 1,000 executives |
| 75% of Greek CEOs say they are ready for AI; 10% saw both revenue growth and cost reduction | PwC 29th CEO Survey, Greek results, via Parapolitika | 62 Greek CEOs |
| €150M scheme for AI aimed at smaller businesses | Development Law scheme, announced February 2026 | A scheme, not a budget line |
| 231 studies of top managers, median response 32% | Cycyota and Harrison, Organizational Research Methods, 2006 | Sponsorship and topic raised response; follow ups did not |
| MIT: 95% of organisations get zero return from generative AI | MIT NANDA, The GenAI Divide, July 2025 | 52 interviews, 153 conference surveys |
| "AI Readiness Index" already in use | ELTRUN and Entersoftone, December 2025; Cisco | |
| "Frontier Firm" | Microsoft Work Trend Index 2025 | |
| ±10 points with 100 answers, ±8 with 150, ±7 with 200 | BORRN calculation | 95% confidence |