Generative AI "Productizes" Human Services
Gen AI offers human-like services that are available on-demand, storable, of higher quality and are more consistent. While that's great news for firms it may or may not be great for humans.
In one Fintech company an AI chatbot replacing human customer service reps cut costs by 95%, reduced resolution time from 45 minutes to sixty seconds and increased median customer satisfaction from 55% to 69%. Examples such as these are why many predict customer service is prime territory Gen AI will automate.
Gen AI does this because it can “productize” human services like customer service. That is, it takes the messy human-to-human interactions and replaces the human providing the service with AI that is always available, always responsive, acts more consistently, and has a wider range of knowledge and better answers than most customer service reps. The same may be true in other service jobs, such as legal services, consulting and copywriting. If such automation happens at scale, this has far reaching implications for businesses, workers and the economy.
Products vs. Services
When I teach my B School courses, I explain the difference between products and services and the pros and cons of each type of offering. While both offer value, they differ in many ways, which have implications on both their production and consumption.
As the table below explains (courtesy of ChatGPT-4o), products and services differ in terms of tangibility, ownership, ability to store, consistency of quality, etc. These differences are crucial for businesspeople to understand.
The pros of products are that you can own them, they can be stored for later use, they’re available on demand, quality and consistency can be ensured, their production can be scaled via capital equipment, and they can be delivered digitally.
Source: Lumen Learning
The pros of services are that they’re highly customizable, allow interaction with the buyer, and they fill niches products cannot. Generative AI will enable the best of both in many ways by productizing human services.
Productizing Human Services
Now Generative AI is not the first product to digitize and thus “productize” services. Music recordings, movies, and TV series have productized musical and theatrical performances, allowing them to be stored, be available on demand, and scaled easily. This has made it possible for artists and actors to hone their recordings to be of the highest quality before release, seen by millions instead of thousands, and make millions of dollars more than just offering performances in front of people in theaters and arenas.
What’s different about Gen AI is that, instead of just productizing celebrities, it will productize the services of ordinary humans while still keeping a number of the benefits of being a service. This makes sense because what AI can do is take the combined experience of humans (often experts) in a particular area, “learns” them and then makes them available to users. AI aggregates, curates, and packages human knowledge and capabilities and offers them 24X7 at low cost.
AI aggregates, curates, and packages human knowledge and capabilities and offers them 24X7 at low cost.
The upshot is that those services will be available on-demand, storable, of higher quality and more consistent, while also being highly customized and less expensive. The company’s payoff will be high but the cost to workers may also be high.
Probably the best example of this was the Boston Consulting Group study of 18 consulting tasks. Given to both well-skilled and less-skilled employees, you can see that both groups improved their performance in certain tasks using AI (nota bene - not in all tasks; in some they performed worse due to overly trusting AI).
You can see that both skill groups improved performance (higher quality) and the gap between the two groups became narrower (more consistency). This will lead to the following outcome; quality improves and consistency increases.
These results are great for consulting firms. However, the ability of AI to enable less skilled workers to close the gap with more skilled workers may lead companies to consider laying off those more experienced (and expensive) consultants for cheaper, less experienced ones. Or just use fewer consultants overall.
The ability of AI to enable less skilled workers to close the gap with more skilled workers may lead companies to consider laying off those more experienced (and expensive) consultants for cheaper, less experienced ones. Or just use fewer consultants overall.
The takeaway for the more skilled consultants is a) the performance gap they gained over time mostly went away and b) they better learn AI or their less skilled peers who use it will surpass them.
Another example: In the article, AI took their jobs. Now they get paid to make it sound human, AI replaced several copywriters. Here is an excerpt explaining what occurred.
“Writer Benjamin Miller – not his real name – was thriving in early 2023. He led a team of more than 60 writers and editors, publishing blog posts and articles to promote a tech company that packages and resells data on everything from real estate to used cars. "It was really engaging work," Miller says, a chance to flex his creativity and collaborate with experts on a variety of subjects. But one day, Miller's manager told him about a new project. "They wanted to use AI to cut down on costs," he says. (Miller signed a non-disclosure agreement, and asked the BBC to withhold his and the company's name.)
A month later, the business introduced an automated system. Miller's manager would plug a headline for an article into an online form, an AI model would generate an outline based on that title, and Miller would get an alert on his computer. Instead of coming up with their own ideas, his writers would create articles around those outlines, and Miller would do a final edit before the stories were published. Miller only had a few months to adapt before he got news of a second layer of automation. Going forward, ChatGPT would write the articles in their entirety, and most of his team was fired.
The few people remaining were left with an even less creative task: editing ChatGPT's subpar text to make it sound more human. By 2024, the company laid off the rest of Miller's team, and he was alone. "All of a sudden I was just doing everyone's job," Miller says. Every day, he'd open the AI-written documents to fix the robot's formulaic mistakes, churning out the work that used to employ dozens of people.”
For the firm, this was a great solution. Lower costs, consistency increased, customizable output, and good quality. But for the workers it’s not been a positive story. Contrary to the idea that AI would take the drudgery out of work and leave humans with the fun and creative tasks, in this example it did the opposite. This is illustrated by this quote from Joanna Maciejewska.
Is there some way firms and workers might both benefit from productizing human services? Moderna might be an example of how that might happen.
Moderna has created over 750 CustomGPTs using OpenAI’s ChatGPT to bring out new products faster. Per the Wall Street Journal, Moderna
“aims to automate nearly every business process at the biotechnology company and boost the ChatGPT maker’s reach into the enterprise.
As part of the transaction, some 3,000 Moderna employees will have access to ChatGPT Enterprise, built on OpenAI’s most advanced language model, GPT-4, by the end of this week. Further integration of AI into more of its processes could help Moderna outpace its plan to roll out 15 new products within the next five years, the Cambridge, Mass., company said.”
Three examples of how Moderna is using CustomGPTs are
Clinical Trial Dose Optimization: One GPT uses years of previous research and medical knowledge to predict the optimal dose of a drug for clinical trials. This addresses the challenge of dose optimization, which is crucial for the success of clinical trials.
Regulatory Responses: Another GPT drafts answers to questions from regulators by combing through swaths of research. This has significantly reduced the time required for this task from weeks to minutes.
Drug Manufacturing: On the manufacturing side, a GPT is used to predict the structure of new enzymes, which can enable manufacturing processes with better yield and reduced waste.
However, depending on your company to deploy Gen AI in a way that benefits you may be risky. A better approach may be to adopt Gen AI yourself to make yourself more productive, knowledgeable and creative.
A (somewhat humorous) example of this comes from this Reddit post:
What’s Next?
As the Fifth Law of technology historian Melvin Kranzberg states,
“All history is relevant, but the history of technology is the most relevant.”
As it seems with much of AI, it will bring mixed blessings and exactly who are the winners and losers is being determined as we speak. However, in the rush to Generative AI, it would be great if those producing it spent more time thinking about its implications for society and humans.
As usual, you nailed it. This is the kind of realistic assessment that can help us all remain disciplined and optimistic.