Posted on

Capitalizing On The Promise of AI

The rapid proliferation of Artificial Intelligence (AI) tools like ChatGPT have created both awe and anxiety among people, and led to intense debates about the future of humanity. How should people prepare for a world where AI has a significant presence? What does the future of work look like?

Business leaders have an equally enormous task ahead of them. What are the threats and opportunities posed by AI? What changes do they need to make to their strategy, operations or culture in order to stay relevant? And how does one handle this radical transformation of work while keeping employees engaged?

One estimate says that AI will unlock $15.7 trillion in productivity gains by 2030. However, companies have to take deliberate steps to adapt their organizations in order to capitalize on the potential that AI offers.   

With Productivity Commodified, Creativity Becomes More Salient

AI is enabling productivity gains that are orders of magnitude higher than what humans can accomplish. AI can spit out code snippets or new graphics in a matter of seconds, compared to hours it takes a person. 

Competing with AI on skills where it can outperform people is pointless. Any business that doesn’t adopt AI productivity and employs people to do the same tasks, will soon fail. Instead, the most promising approach is to harmoniously coexist with AI where people focus on tasks that are uniquely human, like creativity, while AI handles more routine tasks. 

Focusing on innovation is important from a differentiation perspective too. When everyone uses AI to help improve their productivity, productivity becomes commodified. The only way then, to stand out among your peers is by using creativity – finding new ways to use the latest technology or creating new capabilities that delight customers. As an example, the low-code/no-code platforms are allowing people from different industries, including health and finance, to create their own applications and improve their workflows. 

Creativity is where humans have a distinct advantage. AI hasn’t performed well when it comes to creative tasks because creativity, by definition, requires production of ideas that don’t exist before. Even though AI is transforming the creative industry, the nature of its output is only superficially creative. Current set of AI tools are only good at rearranging existing elements in different ways to give an illusion of creativity. When prompted to create something that is truly original, tools like ChatGPT tend to fail. 

Leaders need to recognize that human creativity is now the differentiator for their business and they need to invest heavily in fostering internal innovation. 

Laying Off People Is The Wrong Strategy

Staying ahead of the AI acceleration requires proactively redesigning the boundary between human and AI work. This is a continually shifting line and businesses have to be agile when it comes to rearranging work for maximum productivity and innovation. 

As a business leader, it might be tempting to think that with AI taking over more and more tasks, your company needs fewer people to do the work that was being done before. This mindset can be dangerous for a company’s competitive advantage because it does not tap into the productivity gains that any new disruptive technology like AI unleashes. Barring a few domains (like some service industries), for any industry that has scalability potential and demands continuous innovation, reducing overall headcount will slow progress and almost certainly backfire in the long run. 

To see why this approach is damaging in the long run, let’s assume that your work requires some level of innovation to stay competitive in your market and therefore needs humans-in-the-loop. As more work starts shifting towards AI, it seems reasonable to reduce headcount in order to retain the same level of productivity for a much smaller cost. 

However, after a certain level of reduction, people become the critical resource in an organization. Even assuming that people are only doing tasks related to creativity that AI doesn’t handle well,  the organization becomes limited by the level of innovation it can harness from its employee base. As the AI capability expands even more (at a much faster rate than people upskilling), the extra capacity remains untapped due to limited people resources. After a short-term boost in productivity and lower costs, organizations get hamstrung in their ability to produce outsized innovations that can put them ahead of their competitors. 

Instead of simply reducing headcount to maintain the current levels of productivity, leaders need to take a long-term approach to talent management. Retaining and hiring people with the right skill set for innovation will help organizations take advantage of growing AI capabilities and provide higher levels of productivity and innovation compared to their peers. 

Engineering A High Performance Culture

A natural consequence of the increased productivity offered by AI, is that it allows more complex work to be possible. Complex work is high on both innovation and productivity, and the more complex the work, the more you need to tap into the intelligence and expertise of others. However, handling complexity, when humans are a critical part of the loop, is tricky. Groups can behave as intelligent swarms but they can also arrive at incredibly poor outcomes. How smartly a group behaves depends on the overall ability of individuals in the group as well as how independently they are allowed to think. Relatively smart people when thinking without the influence of others, cancel out each other’s errors and biases leading to much better decision making and problem solving that wouldn’t be possible at an individual level.

The single most important thing that organizations can do, in addition to hiring good talent, is to set up processes, tools and norms that allow people to contribute ideas and participate in decision making in independent ways. Without the right performance-focused culture, organizations will find it hard to capitalize on the higher complexity demands. 

Leadership Takeaways

Any transformative technology, by its very definition, radically changes the way people do things and opens up new markets and opportunities. How companies respond to the new environment determines how well they can capitalize new opportunities. 

Compared to disruptive technologies of the past, AI places new challenges due to the rapid pace of development. Companies need an agile approach to managing people and work. 

  • Innovation Management is key: With high productivity gains that benefit everyone, It’s inevitable that people-work will shift towards creativity. How well companies harness employee innovation will determine how they differentiate in the market and tap into the value that AI provides.
  • Hire for the long-term: Automation eliminates some jobs but typically creates many more new ones. A Deloitte study found that automation in the UK created over 4x more jobs than it eliminated that on average paid more. Instead of laying off people in order to boost short-term efficiency metrics, companies should focus on retaining and retraining employees to handle the higher workloads that are bound to come. 
  • Build a culture of performance: With AI taking over time-consuming tasks, more complex work that involves higher innovation and productivity will now be possible. Complex problem solving relies on individuals with different expertise to work together towards a common goal. Leaders will need to create a high-performance culture that incentivizes both individual expertise and swarm intelligence. 
Posted on

Harnessing Group Intelligence For Innovation Productivity

Edward DeBono argued three decades ago that creativity is the most important human resource of all, and it’s truer now than ever. Without a strong culture of innovation that harnesses employee creativity effectively, organizations are much more likely to fail in the modern economy.  

But capturing employee creativity towards organizational innovation is challenging. Our current management practices evolved from Fordist (and Taylor’s scientific management even earlier) approaches to measuring productivity that served well for linear, predictable systems. For example, if a production line produces X widgets/week, you could be sure that by adding another production line you can get 2X widgets/week. Unfortunately, applying these approaches to innovation doesn’t work for two reasons:

  • Creativity is fundamentally non-linear. During brainstorming two people with individually weak ideas might discover key insights that lead to a billion dollar one. Or, an individual could be wrestling with a challenging problem for several days with very little to show for it and then suddenly find an interesting solution in a day. 
  • Complex problems often require a group to collaborate in order to find efficient solutions. But simply bringing people is not enough. Depending on the group and how individuals interact, you could get collective intelligence or collective stupidity. Most incentives are aligned with individual performance and backfire when work is highly dependent on group performance. 

So instead of focusing on linear metrics like bugs fixed per week or number of features shipped, we need better mechanisms that allow for collective intelligence to flourish. 

Swarm Intelligence

Some of the most fascinating examples of complex problem solving in groups come from nature. Groups of ants, bees, birds and other creatures can solve surprisingly complex problems easily even though each individual in the swarm is only following simple rules. Intelligent behavior emerges from these simple interactions. For example, army ants can travel for long distances without traffic jams in dense three lane highways by following simple rules like outgoing ants turn aside when they encounter incoming ants. 

Collective intelligence works surprisingly well in many kinds of problems like estimation or prediction. One of the most well known examples comes from Francis Galton, a pioneering statistician and half-cousin of Charles Darwin, who asked people to guess the weight of an ox at a fair in 1906. While individual responses varied from 1,074 to 1,293 lbs, the mean came to 1,197 lbs closest to the true value of 1,198 lbs! 

Group intelligence also outperforms individual intelligence in much more complex situations like hiring the right candidate for a role. Google at one point had a fairly complex hiring process where candidates could face up to 25 interviews. After analyzing data on interview performance and subsequent job performance, they were able to simplify the process down to 4 interviews which gave them an 86% confidence level. The interesting thing that came out of their analysis was that none of their managers – regardless of the years of experience under their belt –  were individually good at predicting who would make a good employee. The wisdom of the group of four always outperformed individual prediction! 

Research from several domains shows that for groups to work intelligently, few key conditions need to be met:

  • Individual members must be well-informed about their area to have a better than 50-50 chance of getting the right answer.
  • Cognitive diversity of the group is important. 
  • Individual members have to think independently and not influence each other’s opinions. 
  • Individual members should be unbiased.

Most organizations spend incredible amounts of time trying to find the most qualified candidate with the right set of domain skills during hiring, so the first criteria is often not an issue. The remaining three criteria often get underestimated (or ignored) which limits the amount of innovation companies produce.

Cognitive Diversity

The kind of diversity important in complex problem solving is cognitive diversity which includes knowledge (people skilled in different domains), perspectives (different ways of viewing a problem) and problem solving approaches (different heuristics or ways of generating solutions to the problem).

Cognitive diversity is correlated with diversity based on common factors like gender, race or ethnicity. Companies that prioritize increasing diversity have a better chance of being innovative and real-world results back that up. More diverse and inclusive companies are 1.7 times more likely to be innovation leaders in their industry. Scott Page, complexity scientist and author of The Difference, captured this perfectly when he said, “being different is as important as being good.”

Independent Thinking

Perhaps the most important criteria for harnessing group intelligence is for individuals in a group to think independently and not be swayed by others opinions. Unfortunately, this is also one criteria that most groups don’t pay attention to. When groups get together without adequate preparation, a few people might dominate discussions or people might be overly influenced by their leaders’ opinions. In such situations, decisions quickly fall prey to groupthink. 

To mitigate this, individual members need to think about the problem at hand and commit to their ideas and reasoning before they share with the group. This makes a huge difference in the quality of decision making. Despite the popularity of traditional brainstorming, groups that engage in nominal brainstorming (where people brainstorm individually before sharing with the group) come up with twice as many ideas overall and more original ones. 

Unbiased Thinking

Groups also suffer from biases which include both stereotypical (like gender) or cognitive biases. Hiring a more diverse workforce can mitigate the effects of stereotypical biases but cognitive biases can still remain. Two common cognitive biases are myside and one-sided thinking. 

Myside bias results from people’s inclination to favor arguments that support their opinion while ignoring or minimizing contradictory viewpoints. Organizational cultures where “defending your idea” is deemed an important leadership trait accentuate this bias.

One-sided thinking is our preference for arguments that are one-sided rather than those that offer multiple perspectives. A one-sided solution appears simpler and cleaner, and because it causes less cognitive strain, it’s more persuasive. As a result, leaders are easily swayed by a person who presents a simple argument compared to another who presents a more nuanced view. 

Useful Strategies

Groups can be smart or dumb. It’s easier now to see why groups end up with harmful outcomes in the age of social media where most, if not all, of the conditions get violated. Organizations can avoid some of these pitfalls and create a more innovative culture by implementing the right norms and incentives. Here are a few strategies to consider:

  • Hiring a diverse workforce is necessary but it’s not sufficient. Without implementing good systems where diverse viewpoints are included in decision making, one can’t reap the benefits of diversity. Despite the focus on DEI, inclusion from the perspective of innovation is still an issue among many companies. 
  • Creating the right norms for decision making are important when it comes to ensuring independent and unbiased thinking. The Build, Teardown and Rebuild (BTR) method is one such example that removes biases from decision making in groups and can be implemented relatively easily. 
  • Ensure that your incentives and metrics don’t create the wrong kind motivations when it comes to contributing to groups. Complex problem solving is an emergent property when groups follow certain conditions. It requires individuals to behave in prosocial ways that are often at odds with incentives which encourage individual performance at the expense of group benefit. 

Complex and creative problem solving is a highly social problem. What gets created in any company – code or product or service – is heavily dependent on the people that come together to solve the problem. In such an environment, social engineering is as important (if not more) because how intelligent a group behaves depends on how the group interacts and problem solves. Creating the right processes and incentives can improve a company’s innovation throughput and set them on a competitive path.