Bunq in breach of PSD2 - Its AI knows it, but its staff don't

banjopotato

Registered User
Messages
40
My wife and I recently moved our current and savings accounts from AIB to Bunq. We were attracted by the better interest rates on the savings and the cheaper fees for our use case relative to what AIB was charging us.

We were quite pleased overall with the bank. Their front-line customer service was Artificial Intelligence, which seemed weird at first but the AI was really a lot better for most things than most of the frontline telephone staff I've dealt with at AIB or BofI. It knows its stuff and doesn't have any difficulty understanding questions, etc.

That was all fine until I got notification from the app of 2 transactions for over €400 that I immediately knew were fraudulent. They had been made in the US, online on Amtrak.com (which explained why I was not asked to approve them in the app). Within minutes, I had reported the unauthorised charges to the bank, asked for a chargeback, and blocked the debit card used. I also in the ensuing days reported the fraud at the local Garda station.

Under Article 73 of PSD2, banks are to refund unauthorised transactions to the customer immediately upon having been notified of them and in no case later than 1 business day. The fraud happened on a Sunday. Monday there was no refund of the money made. Nor on Tuesday. On the Wednesday, I made a formal complaint to Bunq pointing this out.

Since then, I've been in touch not only with the AI support but also with human support through Bunq's SOS Hotline and the human support chat. It has been a surreal experience to say the least. The humans are mostly woefully misinformed and have told me a few falsehoods like: "PSD2 only applies to direct debits" and "the one-day delay only applies from the time the MasterCard investigates and makes the determination that the charges were unauthorised." Both claims are nonsense.

Meanwhile, the AI support immediately recognised that the bank was in breach of its regulatory obligations and was deeply apologetic! It's a sad commentary that not only is the AI better informed, but it's also more empathetic than the human staff. The problem is: the AI can't actually do anything and there's no evidence that the human staff listen to it even when it's doing the AI equivalent of raising the alarm.

It's now been 10 days since the unauthorised transactions and I'm no closer to getting a resolution. The staff keep updating with messages with "good news" about how MasterCard is undertaking their investigation and I could have my money back as soon as in 45 days (!). Meanwhile the AI is beside itself with contrition, telling me how unacceptable it is that the bank is not meeting its regulatory obligations and sympathising with me about how frustrating the whole experience must be.
 
Last edited:
I do know that Irish banks are traditionally much more generous and speedy when it comes to customer refunds.

PSD3 is under negotiation and one aspect is to make these things more standard across the EU.
 
The latest hallucination from Bunq (human) staff in the support chat was a concession that Article 73 of PSD2 does require the bank to refund unauthorised transactions promptly but that "in practice" they can take more time. To which I replied: "There is no distinction between what PSD2 requires you to do "in theory" and what you actually must do "in practice." If, in practice, you are not respecting the deadlines described in the regulation, then you are in breach of the regulation. Full stop."
 
Was it a card transaction? While you are correctly summarising the obligations in Article 73 of PSD2, the reality is they will be investigating to see if you were careless or negligent with your payment credentials. Read the subsequent articles in PSD2.
 
Meanwhile, the AI support immediately recognised that the bank was in breach of its regulatory obligations and was deeply apologetic! It's a sad commentary that not only is the AI better informed, but it's also more empathetic than the human staff. The problem is: the AI can't actually do anything and there's no evidence that the human staff listen to it even when it's doing the AI equivalent of raising the alarm.
Fun fact: LLM AI chatbots generally have a bias in favour of giving you the answer you want to hear; it's one of the reasons that they're not reliable.

[And it's one of the reasons why organisations in the business of selling things to people should be slow to use AI chatbots as a customer interface. Yes, you save money on staff, but you end up with customers being told things, or offered prices, that actually you'd rather not stand over. At best, you end up with very pissed-off customers (as here); at worst, a court holds that you're actually bound by what your chatbot said.]
 
Bunq (human) staff in the support chat
I can only assume you are continuing this form of engagement with Bunq because you are enjoying it; as its clearly a waste of time otherwise.

1. You make a formal complaint asking for their final response. From memory they have 21 days.

2. If you are not satisfied with the final response letter you make a complaint to Ombudsman seeking resolution & compensation.

Let us know how you get on.
 
@DannyBoyD

For me and for most people, the preferred way to resolve a problem is to talk to the right person who takes ownership of it and solves it.

I believe that writing a letter means that it goes into the system and could take months to get a reply.

At some stage, you have to make a formal complaint in writing.

It's a matter of choice and opinion when to drop the efforts to solve it by talking to people.
 
Back
Top