'Job satisfaction will go up': How generative AI is changing work

Transcription:

Penny Crosman (00:03):

Welcome to the American Banker Podcast. I'm Penny Crosman. It's a common refrain that in banking and in other industries, AI won't take away jobs. It will just make boring jobs. More interesting to discuss the veracity of this kind of statement. We're talking today with Alenka Grealish, principal analyst at Celent. Welcome Alenka.

Alenka Grealish (00:25):

Thanks, Penny.

Penny Crosman (00:26):

Thanks for coming. So let's start with generative AI. We've been writing a lot about the different ways banks are deploying generative AI. Citi, for instance, has been rolling out GitHub co-pilot to its developers to help them create software that uses reusable code from the bank's repository. Citizens Bank has been creating an AI co-pilot to help call center reps answer questions or other banks doing the same. Ally Financial is using generative AI to summarize calls contact centers. JPMorgan Chase is providing a large language model to all of its employees to help them write emails or email drafts and draft reports. It also uses generative AI to understand and detect phishing attacks. Are any of these kinds of efforts changing jobs today? And if not, are there other projects you're seeing that are changing jobs today?

Alenka Grealish (01:26):

I think what's striking about generative AI is that it was launched now almost two years ago as a consumer product, right? And that's highly unusual. All of a sudden, we as consumers can play with this new at the time, shiny, shiny objects. So that wave has just taken, I think employers by surprise because their young employees were already using Gen AI at work, and so they had to react much faster than they normally would with a tech breakthrough. And the two areas you called out in particular, code development call center are certainly areas where we're seeing a lot of investment energy. I do a leaderboard of announcements and you rattle off several of them and there's a heavy concentration in code and call center, and they're getting results. And these results are helping fuel the next generation of proof of concepts. And by results I'm hearing about just double digit is the baseline, right? And you're often hearing 30 to 50% improvement in productivity for routine coding, up to 70% for manual code reviews in call centers, call summarizations in seconds instead of minutes, and the list goes on. So with these relatively eye-popping productivity gains, banks are getting more confident to explore other areas, primarily still with the productivity lens. But now I'm also seeing the lens on customer engagement. And I'll pause there because I know you've got other questions that'll go down in those directions.

Penny Crosman (03:23):

Well, and just to follow up on that, so the banks are seeing productivity gains in software development and in their call centers for instance. Do you think it's actually making those jobs feel more satisfying, more interesting, leading to company morale or anything like that?

Alenka Grealish (03:46):

Yes, is the short answer because if an employee can do something more interesting, I help that consumer with a more challenging problem. Better and faster. Job satisfaction is going to certainly go up when you don't have to do routine dull parts of a job. And you have earners relationship managers that are earning six figures plus who have boring parts of their job. And if you can minimize those, those managers are going to be much more interested in their job and be able to add value. And so the customer in the end is a winner. I always say one of the underlooked aspects of generative AI is the ability to attract talent. If you've got gen AI tools, that is an asset, and especially for your young generation of employees coming in already ... using these tools, they're going to seek out employers who can turbocharge them in their new jobs and train them better and faster.

Penny Crosman (05:03):

And pay the subscription fees so that they don't have to. In a survey that American Banker did recently, one of the questions we asked was, in what areas is your organization using or likely to use generative AI to support your business over the next 12 to 18 months? About 45% said for general office productivity, 36% said to help fight fraud. 24% said to help developers generate code. 19% said to review or assess client data. 18% said to document compliance, 17% said to create legal document summaries. 16% said for proposal and pitch material and 11% said to recruit or hire new employees. Does that reflect what you're seeing? Does that make sense to you?

Alenka Grealish (05:55):

It does. We've done several generative AI adoption surveys, and the most granular one we did was asked what we consider innovation forward. So those that probably have the broader scope of investment, and we found similar. The overarching theme is around productivity, employee facing, middle back office and functional areas. But we're also seeing energy in customer engagement front office, not only upstream sales prospecting tools, but also you cited email draft generation. That's another use case. And we're seeing other downstream use cases. So I think we're going to see that line item, and I didn't hear it in your list, grow in terms of investment.

Penny Crosman (07:00):

Yes, that definitely makes sense, especially if all the productivity gains you were talking about seem to play out in that use case. Another thing that we asked in our survey was how do you expect AI in all of its current iterations might hurt your business? And there were several different things on the pick list, but 50% of the bankers said skills degradation, for instance, a reduction in critical thinking or analytical skills, and 47% said job losses. So to that first point, what do you think of that idea that people using generative AI or other forms of AI will fail to learn certain basic skills or kind of dumb down the job?

Alenka Grealish (07:53):

I think that that's a real concern. And given that generative AI still can hallucinate produce very confident-sounding wrong answers, the issue really is about how we educate people. And that starts with middle schoolers and assure that they are trained to interrogate gen AI answers and that they look beyond what generative AI can do. My greatest fear is we lose some of our unique human skills, the ability to build on legacy experience and coalesce a lot of information, some of it in print, some of it in voice IE interviews. And something that I always tell my kids when they do a research project is talk to people, interrogate the information you have and allow that to generate more challenging, interesting questions for experts. And I think that's what we risk is not loss of the basic skills, but the human ability to combine print thought and interaction with other humans experts that were at risk of losing.

Penny Crosman (09:26):

Interesting. And what about the point about job losses? So 47% of the bankers we surveyed said that they worried about job losses. Do you think that that's a legitimate worry? And which jobs do you think are most at risk?

Alenka Grealish (09:46):

I don't diminish the fear of job losses. I think that's real. Whether or not that actually will happen I think is open to debate because we've had a variety of tech breakthroughs over the centuries, and there are always some jobs lost, but there's always jobs created. That's why we're still at a relatively low unemployment rate, right? It's around 4%. However, this requires re-skilling and access to education and training that is adapted to new job requirements. In terms of specific jobs, I don't think it's, you could just say wholesale, oh, this job description's at greater risk. I think the people within certain jobs who don't leverage adapt, adopt to generative AI tools, they are at risk of losing their job. But I think there's still going to be people in call centers, how their job evolves will be driven by generative AI. So they're going to become more highly skilled advisers in the next probably 10 years because generative AI will do your basic call center tasks. So how the job evolves will depend on gen AI and how the individual in that job harnesses gen AI will determine whether they keep that job or need to find a different job.

Penny Crosman (11:29):

That's an interesting point. Are you seeing a lot of banks try to re-skill people or help people become prompt engineers and that kind of thing?

Alenka Grealish (11:42):

I not so much yet because the shipped is relatively small right now, but yes, could see how if you've got a person that's very skilled in a call center who just has emotional IQ and really has positive customer feedback, you want to keep that person. So I'm sure they're going to selectively re-skill people who excel at the more human aspects of a job that is now being powered more by generative AI.

Penny Crosman (12:24):

Well, another thing we asked in our survey was what, if anything, is your organization doing to be more AI ready? And only 8% said that they're actively addressing employee concerns about the possibility of job elimination. So I wonder, do you think banks should be having these kinds of conversations, these kind of honest conversations, or do you think it's too early? Because the full impact of generative AI on jobs is not yet clear?

Alenka Grealish (12:54):

I think at the C level, a hundred percent of bankers should be thinking about what generative AI means for their organization, not only on the tech side, but on the organization operational dynamic and the buy-in that that's required. And we found something very interesting when we asked these innovation forward FIs, what the number one key success factor was. We actually asked them what's their top five were, but the number one was buy-in. And I found that heartening because they already recognize that they can get the tech right. They have to invest heavily to get the governance right on the compliance. But the more challenging is how do you get buy-in? And I see a few early movers, best practice players, which first address the risk and fear. You've got to say, Hey, we recognize there's risks here. These AI systems can hallucinate, you can feel threatened that your job's going to get diminished or replaced. Here's how we're thinking about it. And then present the opportunities back to the talent inspiration, retention. Here's what it can do for you and your skills. And lastly, if a company does not think about their gen AI strategy, they risk losing the opportunities that it presents and be an obsolesce, the Kodak moment, how could they rethink their value, add their value proposition in a world of generative AI-powered companies?

Penny Crosman (14:54):

Well, in line with what you were saying a couple minutes ago was in our survey, 76% of bankers said they believe AI will redefine or change the nature of jobs in the industry without eliminating jobs on a large scale. And 36% said they think AI will actually produce new jobs or expanded opportunities. What kinds of new jobs do you think AI might help create?

Alenka Grealish (15:21):

Well, certainly a variety of jobs around generative AI, but first I'd say at the enterprise level, there's going to be teams that are going to grow bigger. I think of governance, AI ethics, those teams are going to become more prominent because there will be a lot more AI, not just generative AI, classic AI. So at the top of the house, ethics, compliance, privacy data governance will be increasingly important. And not only how you use your customer data, but external data that gets pulled into models. And then you've got model governance, right? There's a lot of concern around model collapse. As the inputs to models become more sourced from generative AI, there's a risk of model collapse. So model maintenance is going to be a big job. And then in terms of jobs that will evolve quickly. I think already we have customer experience, u ui, UX designers. Now you're going to need more people who monitor how customers engage with artificial agents, right? Not human agents, and improve on that engagement and improve all the handoff to the human in that experience. And then at the more ground level, there'll be prompt engineering jobs which hadn't existed before, for example, that are important to effectively guiding gen AI models.

Penny Crosman (17:24):

That makes sense. So I read recently on efinancialcareers that JPMorgan Chase has some open listings for AI research, summer interns in London and New York for next summer to participate in research projects on applied AI at bank. And the New York internship pays a salary of up to $155,000. Is that in line with the kinds of salary numbers you are seeing? Are banks really paying top dollar for AI talent?

Alenka Grealish (18:11):

So my view is that's a clever way to preemptively pick out top talent at the top of the funnel over three months and secure them for a full-time job. That's why I think the price adds so high. They wouldn't pay an annualized salary to all those people, but for the summer, they'll cherry-pick who they want to offer a full-time job to. So they're kind of getting ahead of the competition through this process. Am I seeing these six figures? Yes. You look at ML model builders now, jet AI model builders. The cool kids are all these super mathematicians that understand these models and know how to program them. And yeah, I think who's going to lose the most will be the universities where these PhDs won't go into academia. They'll go into industry first because the money's so rich.

Penny Crosman (19:27):

Yes. That's an interesting point for sure. Alenka Grealish, thank you so much for joining us today, and to all of you, thank you for listening to the American Banker podcast. I produced this episode with audio production by Adnan Khan. Special thanks this week to Alenka Grealish at Celent. Rate us, review us and subscribe to our content at www.americanbanker.com/subscribe. For American Banker, I'm Penny Crosman and thanks for listening.