Automation isn’t new. It’s been around for as long as humans. We are always striving to do more with less—to be more efficient. From using livestock to help plow a field, to using waterwheels and windmills to grind grain, to using machines to manufacture cars, humans have developed tools to automate routine tasks and make work easier, safer, and faster.
AI is no different. It’s a tool that can automate tasks and help us work more efficiently. And the claims about its potential are no longer hearsay—researchers are starting to release evidence of the impact of workers using AI. The results are encouraging, and two studies recently caught my eye. Both studies highlight that AI can increase productivity, upskill junior workers faster, and improve sentiment. Each of these findings are worth discussing.
The National Bureau of Economic Research (NBER) studied customer service agents who were given access to a generative AI-based conversational assistant as they chatted with customers. The AI suggested responses and provided links to relevant internal technical documents. Separately, researchers at the Massachusetts Institute of Technology randomly exposed half of their participant pool to ChatGPT (a generative AI chatbot) as they worked on professional writing tasks. The treatment group was allowed to query ChatGPT as much or as little as they wanted to help them complete the task at hand.
Though NBER and MIT took different approaches, both their findings help demonstrate the positive impact of generative AI as an augmentation tool. Now let’s dive into what the studies reveal about the impact of AI.
When it comes to increasing productivity and efficiency, there’s little doubt that AI can help workers do more in less time.
The NBER study of customer service agents found that using generative AI increased agent productivity by 14% on average, with some as high as 22%, as measured by issues resolved per hour. Similarly, MIT found that using generative AI helped professionals develop better content faster: those using generative AI took 40% less time and the quality of their work was 18% higher.
The efficiency gains here aren’t to be overlooked. Not only can more work be done in the same amount of time, but employees can refocus their energy on higher-complexity creative tasks and strategic initiatives.
Part of what sets AI apart from other automation tools is that it can help junior workers—whose jobs might previously have been replaced by automation—learn new skills.
Historically, computers have automated tasks that can be clearly mapped out with explicit rules, such as following steps in a process. Often, the implementation of automation reduces the need for the most junior workers. For example, withdrawing or depositing money at a bank has now been automated by ATMs, and those jobs for young professionals are largely gone.
But AI doesn’t necessarily replace junior workers—it can upskill them. NBER found that at two months of tenure, agents with access to AI performed as well as or better than agents with more than six months of tenure who weren’t using AI. In the same vein, MIT’s study revealed that access to AI enabled less-experienced workers to produce higher-quality work faster, compressing the productivity distribution and lowering worker inequality.
Both studies tell the same story: that AI can help people do their jobs better and faster. That’s because AI helps encode best practices. Typically, expertise carries tacit knowledge, but AI helps level the playing field, disseminating knowledge more quickly and equally. This is a critical point to understand, as it carries major implications for the labor market.
Get Saifr's research → AI insights survey: Adopters, skeptics, and why it matters.
Beyond the productivity numbers, these studies also put forth some very encouraging findings about AI’s impact on sentiment and engagement. NBER found that agents using AI were treated more positively by customers, while MIT found that exposure to AI increased job satisfaction, since participants saw AI as a way to raise the quality of their work while saving time.
These findings could be the result of workers feeling better about the volume and quality of their work. It’s important to not forget about this human aspect. AI doesn’t just help us do better, it can have a positive impact on workplace morale.
NBER and MIT aren’t the only ones capitalizing on AI’s ability to help—Saifr is too. (This isn’t a sales pitch, just a discussion of a use case.)
One of the biggest challenges facing compliance and marketing teams within financial institutions is the compliance review process. It’s a major pain point. If you’re a marketer or content creator, your job is to engage customers, and you want to get your materials out in the market as quickly as possible. Compliance reviews, though necessary, can slow you down.
That’s where Saifr started. How could we help marketing teams publish compliant materials more efficiently, and how could we reduce compliance teams’ burden?
AI can be a powerful solution. We created robust natural language processing (NLP) models by using millions of documents representing over 20 years of work by thousands of marketing and compliance experts in various lines of business. Our AI acts as a grammar check for compliance review, flagging potentially non-compliant content—essentially performing a first pass review in just seconds.
Now, we help firms be more efficient and productive. Plus, our AI explains why content was flagged and suggests less-risky alternatives, helping junior workers learn compliance standards.
The potential use cases for AI seem boundless. As AI systems continue to advance, I think we’ll start to see more industry-specific solutions that could be truly transformative. This is just the beginning. With more time, I believe more studies will support the findings above: that AI can help workers both do their jobs better and feel more positive about them.
It’s critical to note, however, that AI won’t be replacing human ingenuity and expertise. AI is only as good as the human data it learns from, so humans will need to continue to develop quality content.
Want to know how compliance and marketing leaders at U.S. financial institutions are using and thinking about AI? Download our ebook AI insights survey: Adopters, skeptics, and why it matters.
The opinions provided are those of the author and not necessarily those of Fidelity Investments or its affiliates. Fidelity does not assume any duty to update any of the information. The information regarding ChatGPT and other AI tools provided herein is for informational purposes only and is not intended to constitute a recommendation, development, security assessment advice of any kind. Fidelity and any other third parties are independent entities and not affiliated. Mentioning them does not suggest a recommendation or endorsement by Fidelity.
1085817.1.0