Could investing in AI automation actually hurt productivity?

Historically, and in the zeitgeist, adopting AI means automating jobs. For talent-centered organizations, is downsizing our only choice? Is it possible to adjust to new technology while keeping talent at the center? Five years ago VC partner, Surabhi (Ruby) Nimkar , told me a story about this that I can’t forget.?

Facing the potential loss of 100,000 workers due to automation, Accenture built an AI to forecast job-automation risk and recommend reskilling opportunities. Over four years the JobBuddy AI helped retain 300,000 employees . As a human, this was heartening, but also unusual. Why would a business make all this effort to update skills? To paraphrase Accenture's HR leader, Ellyn Shook, if you’re planning to stay in business, you must design and equip your workforce for the future. Using AI to automate tasks will not work without evolving the rest of your business, including the people.?

“I think the thing that organizations really have to start doing, quickly, is move from workforce planning to work planning and really understand what work is going to be done by machines and what work is going to be done by humans, and make sure that you are investing in your people, to understand how to work with the technology. And I think that's how you're going to future-proof your workforce." - Ellyn Shook, 2019

The not-so-secret sauce in her statement: pick the right new skills to train. Building a workforce for the future requires intelligent forecasting of where your firm will add value in the future. Can we adopt technology not only to reduce cost but to add more value? Yes, but only if you engage the company in imagining that value.?

It turns out that most growth is found in expanding value rather than streamlining cost. Former Head of Talent, Joshua Craver , and I recently dug into this topic.?

The paradoxical path to unlock AI: invest in people.

According to economists, over the long run, investing in AI alone yields little or no gains in productivity. Or, to put it proactively, to benefit from AI, you also need to invest in human skill development and organizational redesign. Often, companies try to avoid change by implementing what Beanne and Brynjolfsson called “plug-and-play automation: swap in a piece of machinery for each task a human is currently doing.”?

Here, it’s as if you’re making a self-player piano . You build the entire piano as well as a machine to move its keys. With the Walkman and iPod already behind us, and player piano reduced to novelty, we can see how flawed that form of automation turned out to be.?

Consciously or not, most companies are pursuing the player piano approach to AI. But they might find more value, even in the short term, by augmenting their workforce instead. David Gaynor recently introduced me to a mental game for adopting AI: 50% Faster, Cheaper, Better. Which use is more likely to matter? For fun, I sketched some numbers so you can start to see impact of each approach.

Some principles from this worth noting:?

• To help, AI must be half the price of a human or less. This is harder to achieve than it sounds.
• Here, AI is doing about 50% of a person's tasks, but for most roles, the theoretical possibility is closer to 30% .

Obviously, this is just a ruff illustration. For a deeper take, academic literature also explains why rethinking work makes more sense than automating it.

Three research findings support worker augmentation over plug-and-play automation.?

1. Over our economic history, most economic value has come from inventing new goods, instead of making existing ones cheaper. (Bersnahan & Gordon, 1996)
2. Companies that successfully adopt technology also change their organization in significant ways to reinforce its benefits. (Milgrom & Roberts, 1990)
3. With tech and org changes, many new tasks are invented. Sixty percent of the work people did in 2018 didn’t exist in 1940. (Autor, Salomons, & Seegmiller, 2021)

To start, try building an AI lab.

As I hinted in an earlier episode , the opportunity is to engage talent in rethinking how they can add value. Together, we can uncover opportunities for reinvention that would enable new work for humans and their machines. If you want to talk about how to do this, Brave has just opened up free office hours. We'd like to see more orgs move in this direction and will back this with our time. Sign up here .?

“If I had magical powers to make my HR colleagues understand one thing it would be that the real power of technology is not to eliminate humans from the equation but to elevate humans.” - Ellyn Shook, 2018

References

Autor, D., Chin, C., Salomons, A. M., & Seegmiller, B. (2022). New Frontiers: The Origins and Content of New Work, 1940–2018 (No. w30389). National Bureau of Economic Research.

Beane, M., & Brynjolfsson, E. (2021). Working with robots in a post-pandemic world. MIT Sloan Management Review, 62(2).

Bresnahan, T. F., & Gordon, R. J. (Eds.). (2008). The economics of new goods (Vol. 58). University of Chicago Press.

Brynjolfsson, E., Mitchell, T., & Rock, D. (2018, May). What can machines learn, and what does it mean for occupations and the economy?. In AEA papers and proceedings (Vol. 108, pp. 43-47).

Feloni, R., & Turner, M. (2020, February 13). Accenture’s AI program “Job Buddy” helps retrain workers who’ve lost jobs to automation, and it could be a game-changing model for other companies to follow. Business Insider. https://www.businessinsider.com/training-employees-on-new-skills-and-technology-what-accenture-learned-2019-1?r=US&IR=T ?

Milgrom, P., & Roberts, J. (1990). The economics of modern manufacturing: Technology, strategy, and organization. The American Economic Review, 511-528.

Lorrie Lykins, L. (2018, March 15). Accenture’s CHRO Ellyn Shook on the Future Workforce. Institute for Corporate Productivity (I4cp). https://www.i4cp.com/shesuites/accentures-chro-ellyn-shook-on-the-future-workforce

As mentioned in the podcast video above, here are two studies that associated employee and customer satisfaction. While the strength of this relationship is still debated, literature points to more evidence for employee experience lifting customer experience, but less evidence for the reverse.?

Jeon, H., & Choi, B. (2012). The relationship between employee satisfaction and customer satisfaction. Journal of Services Marketing, 26(5), 332-341.

Pantouvakis, A., & Bouranta, N. (2013). The interrelationship between service features, job satisfaction and customer satisfaction: Evidence from the transport sector. The TQM Journal. 25, no. 2: 186-201.

We Are All Agents | podcast

AI is rapidly shifting how we organize and work. But we can’t just react to technology if we want a human-centered future. We Are All Agents in shaping change. Let’s help technology and talent work better together.

We’ll share new ways of thinking about and working with machines discovered in our own research and from the work of our guests. Our mission is to help talent, leadership, HR, and developers build more human-centered organizations. How we create, assign, and organize work, is all on the table for change. Let’s use this moment to make our work even more valuable for organizations and people alike.?

If you share this mission and would like to collaborate, or have a question you need help with, sign up for office hours . To learn more about us, here’s our website . We look forward to learning about your work as well. And until then, be brave.