Nineteen mobile banking apps in Azerbaijan were measured independently, entirely from public sources. The store picture gives each app's rating on both stores and the average of its written reviews from the last 90 days; bank behaviour gives the reply rate to negative reviews, the delay and the template share; reach gives who can install the app at all. Alongside them, 30,591 public user comments were classified by AI into twelve complaint categories. A rating says how highly something is scored; the behaviour says what the bank does with an unhappy customer. The aim is observation, not exposure.
Emotix·AI & UX studio, Baku·Updated 31 August 2026·EMTX-2026-08-31-IDX-011
Comparative table
Rating and rating count on both stores, plus the mean star of written reviews from the trailing 90 days. That last figure comes out below the lifetime star everywhere, because people who write a review are more motivated than people who only tap a star, so read it as a gap compared across banks rather than as a corrected rating.
1
m10
PashaPay
4.85
245,480
4.36 (-0.49)
4.88
116,121
2
Dostbank *
Bank of Baku
4.84
4,299
3.89 (-0.95)
4.62
1,799
3
Birbank
Kapital Bank
4.82
692,304
4.26 (-0.56)
4.87
191,617
4
PASHA Bank
PAŞA Bank
4.82
40,662
3.78 (-1.04)
4.88
12,000
5
Leobank
Unibank
4.77
142,186
4.02 (-0.75)
4.88
58,973
6
ABB mobile
ABB
4.74
268,828
3.44 (-1.30)
4.85
115,371
7
Express24
Expressbank
4.61
7,870
n/a
4.83
6,132
8
TuranBank Mobilbank *
TuranBank
4.33
2,672
n/a
3.59
111
9
XalqOnline
Xalq Bank
4.20
1,811
n/a
3.16
263
10
atb360 *
Azər-Türk Bank
4.09
48
n/a
4.01
190
11
Premium Mobile Bank *
Premium Bank
4.07
95
n/a
3.98
45
12
UBank
Unibank
4.04
30,089
3.72 (-0.32)
4.81
19,979
13
Rabita Mobile
Rabitabank
3.96
9,801
3.48 (-0.48)
4.45
3,668
14
Yelo *
Yelo Bank
3.95
531
n/a
n/a
n/a
15
BankAvrasiya *
Bank Avrasiya
3.88
97
n/a
4.56
16
16
Mobil Şöbə
Bank Respublika
3.26
3,327
n/a
3.36
425
17
AFB Mobile
AFB Bank
3.23
237
n/a
3.98
113
18
myAccess
AccessBank
3.03
926
n/a
3.36
293
19
Ziraat Bank Azerbaijan *
Ziraat Bank Azərbaycan
2.88
370
n/a
3.66
93
Click a column header to sort. A starred row carries a note, hover to read it. Every figure is taken directly from public store sources on the day the index was captured.
Rating distribution
m10
4.85
Dostbank
4.84
Birbank
4.82
PASHA Bank
4.82
Leobank
4.77
ABB mobile
4.74
Express24
4.61
TuranBank Mobilbank
4.33
XalqOnline
4.20
atb360
4.09
Premium Mobile Bank
4.07
UBank
4.04
Rabita Mobile
3.96
Yelo
3.95
BankAvrasiya
3.88
Mobil Şöbə
3.26
AFB Mobile
3.23
myAccess
3.03
Ziraat Bank Azerbaijan
2.88
Key findings
01
The store star is old news
Ruiz et al. showed in 2016 that the lifetime star displayed in an app store barely responds to anything a team ships, because years of accumulated ratings hold it in place. We measured that in Azerbaijan. Averaging only the written reviews from the trailing 90 days produces a figure below the store star for every single app. ABB mobile shows 4.74 and its recent written reviews average 3.44. PASHA Bank shows 4.82 against 3.78. Dostbank shows 4.84 against 3.89. Because people who write are more motivated than people who only tap a star, this figure comes out lower everywhere, so it should be read as a gap compared across banks rather than as a corrected rating. The narrowest gaps belong to m10 and Birbank, the widest to ABB mobile and PASHA Bank.
02
Replying and answering are not the same thing
Three banks reply to almost every negative review, and each does it in a completely different way. Rabitabank replies to 100% of one- and two-star reviews with a median delay of 117 days, and one reply in ten arrives after 535 days. Birbank also replies to 100% of them, the same day, but only 12% of those replies are distinct text and the rest are one template. Bank Respublika's Mobil Şöbə replies to 98% and 90% of its replies differ from one another. All three could write the same sentence in a report, that we answer our customers, while the customer experiences three different things: an answer after half a year, an instant identical sentence, and an answer somebody actually wrote.
03
Five apps never answer an unhappy customer
Yelo, atb360, Premium Bank, Bank Avrasiya and Ziraat Bank Azerbaijan replied to none of the one- and two-star reviews in the sample. AFB Bank replied to 2%. The cost of that has been measured: across 4,350,541 reviews, Hassan et al. found a user who receives a reply raises their rating afterwards 4.4% of the time against 0.7% for a user who does not, roughly six times the chance. This is a free lever, entirely under the bank's control, and it is sitting unused.
04
The two stores rate the same bank differently
Hu et al. found that at least 79% of cross-platform apps carry inconsistent stars between iOS and Android. In Azerbaijan the gap sometimes exceeds half a star. Xalq Bank's XalqOnline scores 4.20 on Google Play and 3.16 on Apple. TuranBank scores 4.33 against 3.59. In the other direction UBank scores 4.04 on Google Play and 4.81 on Apple. That is why this index never averages the two stores and reports each separately. A bank watching its app on one store only is seeing half the picture.
05
The languages declared on the listing are a finding in themselves
Ziraat Bank Azerbaijan's Apple listing declares English and Turkish only, with no Azerbaijani. Bank Avrasiya declares English only, even though its own app description is written entirely in Azerbaijani. Dostbank declares the opposite, Azerbaijani only, with no Russian or English. TuranBank declares 48 languages including Albanian, Belarusian, Hindi and Japanese, which means the developer never set that list and it stayed at its default. All four cases show the same thing: at many banks the store listing is not treated as part of the product, even though it is where a customer meets the app for the first time.
06
One bank's own website links to an App Store page that does not exist
Yelo Bank's site sends iPhone users to the App Store. That link points at apps.apple.com/az/app/yelobank/id6754243538, and there is no page there. All five storefronts checked return 404, namely Azerbaijan, the United States, Turkey, the United Kingdom and Russia, and Apple's public lookup interface does not recognise the identifier in any country. The bank's Android app works and carries 531 ratings, so the issue is not whether an app exists; it is that the bank's own page hands the customer a dead link. This is one line of markup and a day's work to fix, and it survives because nobody measures it.
07
Whose phone can install it, and whose cannot
The minimum operating-system version an app demands decides who can install it at all, and nobody has measured that in Azerbaijan. Leobank, Rabita Mobile, Mobil Şöbə and AFB Mobile require iOS 16, and m10 requires 15.8. atb360 and Dostbank stay at 14, making them the two most open to older phones. Size varies more sharply still: Birbank is 585 MB, Ziraat 354 MB, Leobank 366 MB and Premium Bank just 23 MB. That is more than a twenty-five-fold spread, and the person who pays for it is the customer downloading over mobile data.
08
The largest improvement in the index belongs to Dostbank
In the June edition Bank of Baku's Dostbank sat last at 3.2, with two in every five reviews at one star. Its Google Play rating is now 4.84 and its rating count has risen from 1,888 to 4,299. That is the largest improvement we have measured since the index began, and it shows a position is not fixed. At the same time its trailing-90-day written reviews average 3.89 and its median reply to a negative review takes 13.4 days, so the work is not finished.
09
Five of twenty-one banks follow the regulator's own rule
Clause 7.3 of the Central Bank's Metodoloji Rəhbərlik requires every bank to publish a half-yearly summary of consumer complaints on its own website and keep it for three years. All twenty-one licensed banks were checked by hand. Five do it: TuranBank, Azər-Türk Bank and Yapı Kredi exactly half-yearly, while Expressbank and Bank of Baku publish quarterly, more often than required. Expressbank's archive reaches back to 2015 and Bank of Baku's to 2016. Premium Bank publishes real statistics but once a year, so it is counted as partial. The remaining fifteen have a complaints section and no summary in it. The rule is followable, and the banks following it are the mid-sized ones rather than the largest.
10
One bank states the obligation on the page where it fails it
Unibank's complaints page quotes the rule verbatim, stating that the consolidated register must be placed in the complaints section of the bank's website every half-year and kept there for at least three years. That same page publishes no such register; the only download is a file of internal procedures. It then adds that summary data is disclosed to consumers only on request. This is not an oversight, because the bank knows the rule well enough to print it.
11
Identity verification is a mid-sized bank's problem
Widening the pain map to sixteen apps surfaced a pattern that was invisible before. The identity and KYC category accounts for 21.6% of comments on Xalq Bank's XalqOnline, 14.4% on AccessBank's myAccess, 11.1% on Bank Respublika's Mobil Şöbə and 10.7% on TuranBank. At the large apps the same figure is 1.7% for Birbank and 2.1% for Leobank. Registration and document verification are effectively solved at the biggest banks and are stopping customers at the door everywhere else. That is not a feature inside the app; it is customer acquisition itself.
12
Complaints concentrate on getting in and staying up
Across more than thirty thousand classified comments, two categories dominate every other. Login and authentication account for 33.3% of comments on Ziraat Bank Azerbaijan, 27.7% on Mobil Şöbə, 26.5% on XalqOnline and 25% on PASHA Bank. App freezes and crashes account for 32.9% on AFB Mobile, 29.6% on PASHA Bank and 27.8% on Ziraat. These are not requests for new features; they are the basic job going undone, where a person either cannot get in or the app dies in their hand. For comparison, m10 sits at 11.2% for login and 7.6% for crashes.
13
The dissatisfaction share varies fivefold between banks
Among sampled reviews, the share at one or two stars is 15% for m10 and 16.5% for Birbank. At the other end sit Ziraat Bank Azerbaijan at 83.3%, AFB Mobile at 69.8%, Mobil Şöbə at 63% and XalqOnline at 55.5%. This measures something different from the overall rating, because the lifetime star carries the past while the dissatisfaction share shows the mood of the people writing today. Where more than half of them are unhappy, the problem is not in individual features but in the whole of the experience.
Market context
Business Insight's “Mobile App World” survey was published in October 2025. The survey is second-hand material from a news source; the primary document should be confirmed separately. The two cuts below line up with this index.
Usage share
Birbank 82%
ABB Mobile 39%
Leobank 28%
Satisfaction
ABB Mobile 94.8%
Leobank 93.9%
Birbank 91.2%
Usage share and satisfaction are different measures. Birbank clearly leads on usage, while it sits third on satisfaction and on store rating. The reply-behaviour section of this index shows the same gap from another angle, because Birbank answers every negative review and answers it the same day, yet only 12% of those replies are distinct text. Scale is one thing, experience is another.
The regulator's own rule: who follows it
Clause 7.3 of the Central Bank of Azerbaijan's Metodoloji Rəhbərlik, 15 March 2022, Protokol 09, requires every financial institution to publish a half-yearly summary of consumer complaints in the complaints section of its own website and keep it there for at least three years. Every other measure in this index is one we chose; this one the regulator wrote itself. Nobody had checked whether it is followed, so all twenty-one licensed banks were checked by hand.
Of 21 licensed banks, 5 publish the complaint summary and 1 publishes it partially.
Every verdict rests on the rendered content of the page, not on an HTTP status, because three of these banks return 200 for URLs that do not exist. For all five compliers the report files themselves were opened and confirmed to be complaint statistics.
User pain map
A rating shows how high something stands, but not where the pain comes from. For this map, the 30,591 public comments across 16 apps (13,611 Azerbaijani, 11,275 Russian, 5,705English) were classified by AI into twelve complaint categories. Click a cell to open the three most resonant comments for that bank and category. Colour shows density only; the severity figure is in the comment panel. Comments stay in their original language with a “See translation” link.
An app needs at least 50 classified reviews to be drawn here. On a small corpus a share is misleading: a topic mentioned in a handful of reviews reads as tens of percent and colours the cell exactly like a genuine thousand-review pattern. The three apps below that line are atb360, Premium Bank and Bank Avrasiya, whose classified corpora run to 6, 19 and 15 reviews respectively. They are named rather than hidden.
Bank
Login & authentication
Failed transfers
Card operations
App freezes & crashes
Customer service
Identity & KYC
Currency & conversion
Fees & pricing
Design & usability
Push notifications
Security & fraud
Update & version regression
m104,300 comments
Dostbank577 comments
Birbank5,627 comments
PASHA Bank1,257 comments
Leobank3,770 comments
ABB mobile5,197 comments
Express24607 comments
TuranBank Mobilbank373 comments
XalqOnline555 comments
UBank4,368 comments
Rabita Mobile2,098 comments
Yelo180 comments
Mobil Şöbə1,229 comments
AFB Mobile73 comments
myAccess250 comments
Ziraat Bank Azerbaijan90 comments
Each cell's percentage shows how often that category appears across all of that bank's comments. Colour intensity is the relative level of that percentage.
Ratings. Read directly from both stores on 31 August 2026. The Google Play figure comes from the listing's own ld+json block, the Apple figure from the public iTunes lookup for the country=az storefront. The two stores are never averaged.
Trailing-90-day average. Computed from the review stream itself, suppressed below 20 reviews. Because writers are more motivated than star-tappers it runs lower everywhere, so read it as a gap across banks.
Reply behaviour. From the newest 200 Google Play reviews, the one- and two-star ones are separated and measured for whether a reply came, how many days it took, and whether the reply is a repeated template.
Reach. Minimum iOS version, download size and declared interface languages, taken from the Apple listing.
Package verification. Every Google Play package name was read back out of the listing and matched to its label, because landing on a similarly named foreign app is easy.
Comment mining. Public AZ/RU/EN comments from app stores and public social channels; short and emoji-only comments dropped. The remaining 30,591 comments AI-classified into 12 categories, multi-label (0-3 severity).
Limitations
Behaviour metrics come from the newest 200 reviews per app. Where an app's review stream is quiet those 200 reviews span years, so the sample-span column must be read alongside them.
atb360, Premium Bank, Bank Avrasiya and Ziraat Bank Azerbaijan have samples too small for a statistical read. They stay in the table because we will not hide who was left out, but their figures are not for comparison.
The trailing-90-day average counts written reviews only, not silent star taps, so read the gap between banks rather than the absolute level.
Declared languages are what the store listing states. We make no claim that every one of them works inside the app.
Yelo Bank's app was not found in the Azerbaijan Apple storefront through the public lookup interface on the measurement day; that row needs a manual check.
AI classification works with a margin of error; read only the overall cell signal, not the label on any single comment.
Each publishing cycle re-pulls the numbers and stamps the date.
Who is not in the index
Four licensed banks have no retail app in the table yet: Bank VTB Azərbaycan, Yapı Kredi Bank Azərbaycan, Azərbaycan Sənaye Bankı and the Baku branch of Bank Melli Iran. We name who is left out on purpose, because the commercial benchmarks this index competes with publish neither their inclusion threshold nor their exclusions.
Frequently asked questions
Which banking app is the best in Azerbaijan?
In Emotix's 2026 banking-app experience index the highest Google Play rating belongs to m10 (4.85), followed by Dostbank (4.84), then Birbank and PASHA Bank (both 4.82). The index also shows the store star is not the whole picture: for every app, written reviews from the trailing 90 days average below the lifetime star. The edition covers nineteen apps as of 31 August 2026.
Which Azerbaijani banking app has the lowest rating?
The lowest Google Play rating is Ziraat Bank Azerbaijan (2.88), where 83.3% of sampled reviews sit at one or two stars. AccessBank's myAccess (3.03), AFB Mobile (3.23) and Bank Respublika's Mobil Şöbə (3.26) follow.
Do Azerbaijani banks reply to complaints in the app stores?
The spread is enormous. Birbank, ABB mobile and Rabita Mobile reply to 100% of one- and two-star reviews. Yelo, atb360, Premium Bank, Bank Avrasiya and Ziraat Bank Azerbaijan replied to none of them. Speed is a separate measure: Birbank answers the same day, while Rabitabank answers everything but takes a median of 117 days to do it.
What is the Azerbaijan banking app experience index?
It is Emotix's independent study of Azerbaijani banking apps built entirely from public sources. It measures three things: the store picture, meaning ratings on both stores plus the mean of written reviews from the last 90 days; behaviour, meaning reply rate to negative reviews, reply latency and how much of each reply is a pasted template; and reach, meaning minimum OS version, download size and declared interface languages. The full method is published so any reader can re-derive the figures.
How many banking apps were analysed?
Nineteen in the current edition. For each app the newest 200 Google Play reviews were read, the one- and two-star reviews separated out, and the bank's reply behaviour measured on those. Apps whose sample is too thin to read statistically stay in the table but carry a note.
Do Azerbaijani banks publish their complaint statistics?
Clause 7.3 of the Central Bank's Metodoloji Rəhbərlik requires every bank to publish a half-yearly summary of consumer complaints on its own website and keep it for three years. All twenty-one licensed banks were checked and five do it: TuranBank, Azər-Türk Bank, Yapı Kredi, Expressbank and Bank of Baku. The last two publish quarterly, more often than required. Premium Bank publishes real statistics but only once a year. The remaining fifteen publish no summary at all.
Why does the store rating differ from the recent-review average?
The star shown in the store accumulates over the app's whole life and, as Ruiz et al. showed in 2016, barely responds to new releases. The trailing-90-day written-review average reflects today. Because people who write are more motivated than people who only tap a star, this second figure comes out lower everywhere, so it should be read as a gap compared across banks rather than as a corrected rating.
Want the full UX teardown of your own app?
This index works on public signals only. An app's real gaps aren't measured without internal data, real user research and a consistent UX approach. Emotix builds this as an ongoing practice for mass-market app experience.
Emotix · AI & UX studio
AI and UX studio. Built in Baku. Independent research, product design and ongoing consulting for banking, fintech, e-commerce and regulated-industry teams building mass-market apps.