CEO replaces 90% of support staff with AI, praises the system on Twitter

midian182

Posts: 9,745   +121
Staff member
A hot potato: Few people are happy about AI taking human jobs. So, if you're a CEO who decides to lay off 90% of a support team, it's probably not a good idea to post a celebratory tweet thread about the "tough" but "necessary" move. That's what Suumit Shah did, and the public response is about what you'd expect.

Shah, the 31-year-old CEO and founder of Bengaluru-based Dukaan, which helps merchants to set up online stores and sell products digitally, posted that "We had to layoff 90% of our support team because of this AI chatbot. Tough? Yes. Necessary? Absolutely."

Several more tweets followed, all of them praising the AI assistant that can answer customer queries instantly and precisely – and doesn't receive any wages or sick leave, obviously. He goes so far as comparing the AI's abilities to Edward Morra from the movie Limitless running the entire customer support section. Again, though, the fictional character of Morra would still want things like money for his services.

Shah adds that the AI chatbot took under two minutes to respond to customer queries, whereas human staff took two hours, and that replacing the team has reduced support costs by 85%.

Throughout the many, many tweets, there is no mention whatsoever of the 23 people laid off, what benefits they received, etc. When someone mentions this fact, Shah seems quite annoyed at the person putting a downer on his AI love fest. "As expected, 'someone' will get offended on behalf of 'someone else,'" he writes, adding that he would post about assistance for his staff on LinkedIn because on Twitter, people are in search of "profitability and not sympathy." There's also a little broken heart emoji, which will doubtlessly please those he recently made unemployed.

"Maybe it was the right decision for the business, but it shouldn't have turned into a celebratory/marketing thread about it," said another user.

Saha told Insider that he regretted posting about the layoffs on Twitter but was adamant that his point still stands. "AI is taking our jobs," he said. "Over time, everybody will start doing this. It's not just us. Maybe I'm just too straightforward to have put it on Twitter," Shah added.

Another organization to replace its (entire) helpline staff with an AI chatbot was the US National Eating Disorder Association. It was quickly disabled after giving out harmful information, including telling a woman to count calories, weigh, and measure herself weekly

Companies such as IBM have said they intend to stop hiring for jobs that could potentially be performed by generative artificial intelligence, which is expected to impact 300 million full-time jobs globally. Stress and anxiety over job losses caused by AI have led to 80% of tech workers using medications, either under a doctor's supervision or otherwise, as coping a mechanism. Heavy alcohol consumption is also becoming commonplace. Still, think of all the money it saves.

Permalink to story.

 
There is a saying: "Somebody is not dumb enough if he is not also bragging about".
He will found quite soon that the next which will be replaced by AI will be himself. Just to maximize the profit to the top.
I'll invest in any AI startup which will design an AI to replace CEO's.
Tough? Yes. Necessary? Absolutely.
 
If the 133 minute to 3 minute 12 second resolution time includes the same or higher success rate, then this seems like an absolute win. IDK how it took that long for humans to answer questions if they are basic enough this AI can do them this fast. Based on my last few tech support calls I may as well have had an AI help me. Frankly, if I'm a merchant, and my support calls get cut down to 2% of their normal time thanks to AI, then bye bye meat sacks! Time is money, and nobody wants to spend 2 hours on a call trying to get answers from someone who cant troubleshoot their way out of a paper bag.

And frankly, he's right. Much like twitter was the first to lay off staff and increase API costs, there will be many companies that can run much more efficiently using this AI, and this will continue to happen. For low level support this make sense, human help is usually only necessary when something cant be fixed easily, which is maybe 1% of calls.

"Another organization to replace its (entire) helpline staff with an AI chatbot was the US National Eating Disorder Association. It was quickly disabled after giving out harmful information, including telling a woman to count calories, weigh, and measure herself weekly"

Amazing what AI can do when not drug down by progressive dogma.
 
Sounds like the CEO needs to be replaced with AI and remove all pay and benefits he has accumulated over the years, or better yet, have the Govt. step in and tax them the combined salaries & benefits of those removed and feed it directly to the former employee's. That would be a good approach to ALL companies to insure employee's and former employee's are able to maintain a satisfactory life.
 
There is a saying: "Somebody is not dumb enough if he is not also bragging about".
He will found quite soon that the next which will be replaced by AI will be himself. Just to maximize the profit to the top.
I'll invest in any AI startup which will design an AI to replace CEO's.
Tough? Yes. Necessary? Absolutely.

What about the other 10%? Can they be laid off too? How about customers? Can they be replaced with chatbots?

This AI trend is the dumbest thing since the dotcom bubble. It will burst soon.
 
If the 133 minute to 3 minute 12 second resolution time includes the same or higher success rate, then this seems like an absolute win. IDK how it took that long for humans to answer questions if they are basic enough this AI can do them this fast. Based on my last few tech support calls I may as well have had an AI help me. Frankly, if I'm a merchant, and my support calls get cut down to 2% of their normal time thanks to AI, then bye bye meat sacks! Time is money, and nobody wants to spend 2 hours on a call trying to get answers from someone who cant troubleshoot their way out of a paper bag.

And frankly, he's right. Much like twitter was the first to lay off staff and increase API costs, there will be many companies that can run much more efficiently using this AI, and this will continue to happen. For low level support this make sense, human help is usually only necessary when something cant be fixed easily, which is maybe 1% of calls.


This thought process is exactly how every customer service line makes you talk to a voice chatbot for 5 mins before "ok let me connect you to a live customer service rep".

 
This thought process is exactly how every customer service line makes you talk to a voice chatbot for 5 mins before "ok let me connect you to a live customer service rep".
The exact reason those chatbots are used is because 90% of calls can be handled without humans.
What about the other 10%? Can they be laid off too?
No. There will always be situations a chatbot cannot handle, that is what this remaining 10% are there for. High level tech support will remain human for a long time.
How about customers? Can they be replaced with chatbots?

This AI trend is the dumbest thing since the dotcom bubble. It will burst soon.
This is just being histrionic. that being said, I WISH my customers were replaced with chatbots. Chatbots wouldnt get so aggressive or mean over basic questions, nor expect me to read their mind.
 
Last edited:
The only relevant question here is what is the quality of the AI-provided support.

If it meets the needs of the end user and the business, then of course it makes no sense to waste money and human time on an already solved problem. Post-transition, the ex-employee and the economy overall will be better off with them doing something more useful. Helping large numbers of society members make this transition simultaneously may take some doing though.

Of course on the other hand if it is not actually providing useful support than it's just another point of user frustration that could have been provided even more cheaply by simply directing users to a web site or community forum.
 
The problem I have with the automated systems, where you say what you want, then get an "ok, try this" response, is when they don't work and switch you to a live person, you have to start the flipping process all over again, so what's the use.
As for live people, the times I've talked to them...A. you can't understand them, B. you can tell they know ONLY what pops up on their screen.
 
The only relevant question here is what is the quality of the AI-provided support.

Have you ever spoken to a support chatbot and got a useful response from one? In my experience they will churn out obvious answers to obvious questions but never actually help you with anything you couldn't have worked out in ten minutes yourself. When they run out of bullsh1t, if you are lucky they redirect you to a real person, more often they tell you 'sorry I cant help, goodbye'. They are awful.
 
Have you ever spoken to a support chatbot and got a useful response from one? In my experience they will churn out obvious answers to obvious questions but never actually help you with anything you couldn't have worked out in ten minutes yourself. When they run out of bullsh1t, if you are lucky they redirect you to a real person, more often they tell you 'sorry I cant help, goodbye'. They are awful.
That's been my experience as well. Although to be fair, some callers are asking obvious questions, and have not spent any minutes working out answers for themselves, so it may be that current or future evolutions will add value in at least those cases. Or if an AI bot could simply more quickly determine that my particular called required immediate escalation, that'd be useful on its own.

Either way, my original comment was just making the point that neither humans nor AI are inherently good or bad just because they are human or AI; it comes down to the details of the particular human / AI and the context it's operating within.
 
Bungalwhere.... This sounds like some mickymouse company in india trying to make a quick buck, no doubt selling snake oil and will crash and burn after this, yada, yada, yawn...
 
That's been my experience as well. Although to be fair, some callers are asking obvious questions, and have not spent any minutes working out answers for themselves, so it may be that current or future evolutions will add value in at least those cases. Or if an AI bot could simply more quickly determine that my particular called required immediate escalation, that'd be useful on its own.

Either way, my original comment was just making the point that neither humans nor AI are inherently good or bad just because they are human or AI; it comes down to the details of the particular human / AI and the context it's operating within.
Wasn't having a go at you about them, I agree with what you said. I was just curious how you/others felt about them and whether your experiences echo mine.
:)
 
What about the other 10%? Can they be laid off too? How about customers? Can they be replaced with chatbots?

This AI trend is the dumbest thing since the dotcom bubble. It will burst soon.
I can't agree less. Why would it burst? What exactly is there to burst? If customer issues were being solved in 3 minutes instead of 2 hours, I don't see an issue here. Anyone can be laid off, no one's job is safe. The days of working the same job for 20 years is likely over for a lot of people.

I've not worked more than 6 years at the same job until the environment, pay and benefits changed in a bad way to the point I couldn't justify staying anymore. Too many businesses don't care enough about their employees to encourage their employees to stay with the business.

People, who's jobs can be replaced by AI, should be training for a career change since AI is only going to improve.
 
Bungalwhere.... This sounds like some mickymouse company in india trying to make a quick buck, no doubt selling snake oil and will crash and burn after this, yada, yada, yawn...
Why would they crash and burn if their customers are happy?
 
I think the real problem is that modern tech support is mostly about following scripts anyway, which doesn't require human beings. Indeed, talking to live humans has usually been just like conversing with bots, which is why I call tech support only as a last resort (and mainly because only they may have some permissions or access I lack). That said, I believe just a touch of empathy for the fired employees would have been in order, as opposed to celebrating the increased company profits as people go jobless.
 
Back