Bank of America Bot Cares About You | Hacker News Reader

Bank of America Bot Cares About You

eksith.wordpress.com

136 pointseksith47 comments

I think those were real, un-empowered humans using a script. My favourite? The "I would feel that way too" script that Google Play uses: straight from an email, no joke, I once got "Thank you for contacting the Google Play Support team. I understand you are waiting on a response. If I was feeling ignored, I would be upset as well. I assure you I have the information for you on your return." or the variant "Thank you for your response. I understand you are having trouble with the tracking number. If I was not receiving necessary information, I would be upset as well. I assure you I have the solution for this brief hang up." (and says there is no tracking number for the item I sent in 3 weeks ago to return! So I replied I was being ignored and got that little gem above...)

I think my favourite has to be the "nothing to say here..." email which emphasizes how they should really cut down the customer service drone talk to the simple sentence "We're looking into it, I'll get back to you." Instead:

Thank you for contacting the Google Play Support team. I understand you are having trouble with your return. If I was experiencing return issues, I would be upset as well. I assure you I have the information to confirm the movement of your refund.

In order to get this issue resolved, I am going to go ahead and let you know that this matter is being consulted, and I will send a follow up email when I receive the information pertaining to your return.

Cordially,

X, Google Play Support Team

Sometimes they forget to personalize entirely, as in the one sent to kick off the botched return process: (the following is unedited)

Hello ,

Thank you for contacting Google concerning your Nexus _. My name is Æsa and I will be taking over your case from this point forward. I understand your concern and why this is important to you and will do the best I can to assist you.

I agree.

Look at this ridiculous conversation I had with Norton support on Twitter the other day: https://twitter.com/esd/status/352070221227503616

I'm sure they're real people working from a script, but they might as well be bot.

Wow. That's nearly as awesome. I mean, it installed something when you uninstall, that's the problem. "We are sorry for the trouble. Do you want to uninstall the Norton tool bar? It was meant to keep you protected. -Arun cc:esd" could be taken both ways, however -- and that's my problem with tweets. Sometimes what we interpret as "clueless" is really someone just "making sure you're okay" using the given script after shooting off too quick with a support response. Not saying they read your tweet, as obviously they didn't respond appropriately about the uninstall, but the responses are consistent with the language on their site: http://safeweb.norton.com/lite (The title says "Is this website safe" and later in the page says, "It's our way of giving back to the online community.")

By comparison, emails are longer-form and less "instant" than tweets. They (Google) should have been able to get it as human as Apple did when I had unactivated iTunes gift cards. (The second time, that is. The first time took them awhile, so it's hit or miss in my experience. I've also learned to never expect real support from Microsoft Store or Adobe -- basically anyone with a "live chat" support service.)

According to https://twitter.com/BofA_Help# -

"Bank of America ‏@BofA_Help 26 May In observance with the Memorial holiday, our Social Media Team will be off Monday, May 27, 2013 and returning on Tuesday, May 28, 2013."

Good to know the machine gets to take a break from its hard work now and then.

Maybe their Social Media Team never came back.

There are some circumstances where it is best to use technology as a tool to empower humans, instead of as a replacement for humans. Customer service is probably one of those circumstances.

I wish Douglas Adams was still alive to see this

I wish Douglas Adams was still alive and writing.

And, someone should make marvin_bot (https://twitter.com/marvin_bot) start-up a conversation with @bankofamerica.

I have a friend that sends these out for bank of America, he works in the office of the president of the company. The company has been ahead of the curve contacting twitter complainers and finding a resolution. These form comments are actually submitted by a person, attempting to keep boa positive as an aggregate of total tweets with mention. No bot included.

Doesn't that stunningly miss the point of sentiment analysis and customer service? If you are optimizing the score with your own output in what way does the sentiment represent a proxy for any meaningful business measure like retention, loyalty or willingness to pay? Surely it's just a proxy for "we have a fella who tweets"?

Truthfully -- I respect a company that doesn't even try to offer good customer service much more than one that tries to fake it

It is a weird paradox, isn't it? Almost as if you don't feel as upset (depending on severity) if a dog bites you than if a squirrel did.

Well, dogs just bite; that's what they do. But squirrels pretend to be all cute and vulnerable and then bite you anyway if you get too close. In your head, you'd be "WTF SQUIRREL?! I TRUSTED YOU, YOU CUTE BASTARD! WHY CAN'T YOU ALWAYS SHOW YOUR TRUE COLORS?"

Man, I hate squirrels. I love squirrels too.

This is an actual thing:

http://www.ncbi.nlm.nih.gov/m/pubmed/14717826/

I wonder if this falls under the category of uncanny valley.

I'm sure it does in some aspect.

A little too friendly and attempting to be helpful in terms of being "human-like", but approaching "not really" in terms of having no emotional feedback and constant repetition (we dislike predictability when negatives are affirmed) is rather upsetting.

Forget the valley, I think eksith is under the influence. If that ain't stoner talk, I don't know what is...

I don't do drugs (except coffee). It's 3:43AM. I tried sleeping for the last 3 hours. Phone rang twice in that time and I'm tired.

It was a joke man. It isn't hard to imagine the guys from that 70's show sitting in the basement and Kelso going off on that squirrel rant.

Hmm. But it's hard to tell what's fake and what's "good" :)

I know it isn't the point of the article, but I can't understand people getting upset at a bank for foreclosing a house. Pay your mortgage and they are unlikely to foreclose.

Except when this happens : https://www.google.com/search?q=bank+of+america+wrongful+for...

You're right in that paying a mortgage will make foreclosure "unlikely", but it's not impossible. Also note, BofA was among many predatory lenders. I.E. those institutions that have outright lied, misled or otherwise duped borrowers into mortgage deals that were woefully unfit for their financial situation. They continued this practice in renegotiations for loan modification after the market crash.

This was revealed by several BofA whistleblowers: https://www.google.com/search?q=bank+of+america+whistleblowe...

Which makes payments to prevent foreclosure unlikely as well.

That's why I said unlikely. There were some issues, and those are unfortunate, but most people had it coming. As for the "predatory" lending, I still have no sympathy. If you are taking out a loan for hundreds of thousands of dollars, you should be doing your own due diligence. It's stupid of them to give those loans because they will lose money, but the guy taking that loan still needs to know if he can afford it or not.

Due diligence is impossible when being lied to, isn't it? I don't think you know what Predatory lending really means.

A lot of people saying, "well they should have looked into it" etc... haven't met anyone who's fallen victim to it and, from my experience, haven't had to face true adversity. This isn't a knock on you, I don't know you. This just from meeting people face to face who say "well it wouldn't happen to me. I've got common sense!" When the piece of paper presented to you is delivered with a smile and a lie (in many cases, half the contract missing), it's not just a matter of diligence.

You're not above human fallibility.

http://en.wikipedia.org/wiki/Predatory_lending

>Due diligence is impossible when being lied to, isn't it? I don't think you know what Predatory lending really means.

Nothing in your article says anything about lying. The main idea is - "imposing unfair and abusive loan terms on borrowers."

The main practices included adjustable rate mortgages that relied on housing prices to continue going up. People took out loans that they could just barely afford at the low intro rate hoping to either sell or refinance after the prices went up.

Other practices included simply offering bigger mortgages than people could afford.

Basically the banks and the borrowers were both greedy and wrongly assumed that house prices would continue to go up. When they didn't, both parties got screwed. The borrowers got foreclosed, and the banks lost a lot of money on the houses they got back.

Lots of people lost a lot of money in the housing crisis (me included sadly), but that doesn't mean they were victims, and it doesn't mean they had to be foreclosed on. If you lost your job and couldn't find another one, that's one thing. If you simply couldn't afford the loan you took out, that's on you. I know the loan docs include lots of pages, but when you are borrowing hundreds of thousands of dollars reading them is probably a good idea. I'm pretty sure even the predatory lending docs still included the % interest, amount borrowed, monthly payments, etc.

If you pay your mortgage on time, why would they ever foreclose?

What's the "^sa" that it keeps appending at the end?

Generally, it's the initials of the Support rep who's replying to the tweet.

I suppose this is all off-topic, but is that common syntax? I've never seen it before.

It was introduced by third-party applications which give multiple people access to a single (or more than one) Twitter account(s). I'm not sure if it was the first to do it, but CoTweet was the earliest that I recall, it would tag tweets with ^CC (initials) based on which CoTweet user account was replying.

Since then it's become fairly standard in customer support, presumably still largely through these third-party apps, but potentially typed manually by some companies as well.

Yep, cotweet was the first app that I saw which seemed to adopt that style

I would say it's extremely common, all Twitter based customer support channels I've ever used have signed off with it.

I thought it was line noise. After all, how does a bot have initials?

Uh they fake them? It looks less obvious this way as when they end with ^unhelpfulbot.

@XboxSupport does that too.

Support Automaton?

I imagine they would have had a much better response if they had just made the twitter handle @BofA_HelpBot

I think a support team is supposed to work with the bot and the bot only initiates contact that the team can then follow up on. Or at least, that's what I'd hope. It seems the support team was afk during that conversation though...

Bank Of America is the worst bank I've ever worked with, just dropped them. We used them for multiple business accounts and their ineptitude is legendary. Stay away. I literally had reps that lied to us about things, left out pages of contracts etc. They were fired but its an epidemic at that place. They forgot who they work for.

Just having a look through the current conversations on @BofA_Help, I suspect that if "SA" is a bot then it's employed to triage extreme negative sentiment; if there's an actual complaint behind it all then a real agent steps in, eg. https://twitter.com/BofA_Help/status/353641709789388800

Quite an efficient way for a customer service team to work, but I'm not sure how it's going to affect the perception of them now that people have pointed it out.

Yep. I bet the "SA" even stands for "Sentiment Analysis."

Scale is the problem. Let's say this wasn't about Bank of America but rather "your favorite giant company" -- the one that has 40 million online users and 16 million mobile users. How does a relationship that large engage in a meaningful, authentic dialogue in tweet-time?

Personally I'd rather they just didn't try instead of faking it. The latter is disingenuous and belittles the customer.

This is great. Someone obviously asked "What's our social media strategy" and got sold some "AI" by someone who said "It's just like SIRI, only better!" and this is what they bought. Sigh.

Well, to be frank, Siri is equally good at pretending to help, without really helping. At least the bofa boat does not say anything in that creepy monotone.

So I guess this is one mark on the "failed Turing Test" side.

oh this could be interesting, a bunch of IT people arguing about automation being wrong :)

I get that you're kidding, but no automation is better than automation that doesn't solve a problem.

Reply on news.ycombinator.com