10 Tips for Leading in the Age of AI
As AI moves from boardroom consideration to implementation, it’s important for leaders to recognise that using AI presents much more than a technology challenge — it presents a leadership challenge.
Artificial intelligence (AI) is a hot topic in boardrooms as executives seek to improve profitability, monetise data, and capture new market share. And as AI moves from boardroom consideration to implementation, it’s important for leaders to recognise that using AI presents much more than a technology challenge — it presents a leadership challenge.
The McKinsey Global Institute estimates that approximately half of the work people do today “has the potential to be automated by adapting currently demonstrated technology.” Note the use of the word “work” and not “jobs”. Many of us at all levels will find ourselves working with machines in our roles, and leading teams of humans and machines.
How will leaders inspire, manage and safeguard the workforce of the future? What new risks and responsibilities will they face? What will agility look like?
Based on our work with hundreds of executives, as well as leading our own organisation as it has grown, we’ve identified 10 actions for successfully leading in the age of AI.
1. Think holistically about opportunities across your value chain, and up and down your P&L
AI opens a new set of performance opportunities. Companies are finding new sources of value from data assets — both from monetising existing data assets and from leveraging external data assets to deliver new capabilities and services to customers. Leaders who think holistically across the value chain, continually seeking out new opportunities, are being rewarded with new revenue streams and increased profitability.
And those that can do this with an open mind toward intra- or inter-industry collaborations can reap even greater rewards. Innovative collaborations, such as data collaborations among retailers, telcos and banks, are elevating customer experiences and helping companies stay relevant in a rapidly transforming digital landscape.
2. Rethink your definition of ROI–and think like a VC
Often, organisations select and measure investments based on the potential return. However, focusing only on financial returns can hinder your company’s growth. In the age of AI, transformative value will often emerge through experimentation. Not every AI project is going to work, nor should it be expected to. But the lessons learned from failures can help advance AI innovation and success.
Formula 1 racing teams, for example, often pursue thousands of ideas and R&D projects, knowing that while the majority may fail, the knowledge that comes with failure will enable them to build a faster race car.
Leaders that create an environment that welcomes exploration, and think of their experiments and associated ROI like a VC investor would, as a portfolio, will innovate, learn and succeed faster.
3. Make diversity a priority
Making diversity a priority was a key theme throughout many of the sessions at SXSW this year. Diversity, including and going beyond gender and racial diversity, is imperative to the success of AI-driven products. It’s also been linked to the ability to attract top talent and greater profitability. (Watch this space for a full blog post on this topic.)
Data science is complex and relies on the interplay of people, cultures, and processes. Gender, racial, cultural, and age diversity can help prevent the algorithmic biases that can derail products and damage brands. And while diversity in your people tops the list, leaders should encourage including diversity in:
• Data sets for more accurate insights
• Data science methods, because hybrid techniques often perform better than individual methods alone
• Academic and disciplinary backgrounds — successful AI projects require not just data scientists, but also data engineers, data translators, domain experts, and change experts.
4. Over-communicate the change
Transitioning to a data-driven organisation is as much a cultural change as a technological one. It’s not uncommon to see employee adoption of new analytics tools lag due to mistrust in “black boxes” or employee morale affected over the fear of losing their jobs to robots.
Successful leaders will offer a clear vision and direct communication about the change — why AI is important to the company’s goals, how it can augment their work, and what every employee’s role will be. By consistently and repeatedly communicating the change, and engaging employees in the conversation, leaders can improve employee engagement and inspire greater productivity.
5. Empower decision-making across the organisation
Leaders have long debated the pros and cons of centralised, top-down management styles and decentralised, bottom-up management styles. In the age of AI, the best approach will be to combine them.
Technology is changing at hyperspeed, making traditional hierarchical approaches ineffective. Simultaneously, the cultural changes and social implications of AI require a controlled framework and strategic vision. As a result, leaders should view themselves as master planners or architects, creating that singular vision and plan while empowering decision-making across the organisation. We think the phrase “greenhouse architect” works well here as, ultimately, you’re designing systems within which individual efforts can flourish.
6. Really invest in understanding the technology
We don’t believe machines will replace humans. Rather they will augment our intellectual reach, just as machines augmented humans’ physical capacity during the Industrial Revolution.
How do you create an organisation that maximises the potential of both its employees and its machines? To answer this question, you will need to understand how AI works and what its limitations are — in other words, what humans are better at than machines and vice versa.
This doesn’t happen without effort, especially for leaders who didn’t grow up through the technical side of the business. Put in the effort — block time in your diary (calendar for some) to learn more about AI, get a reverse mentor, and get technical.
7. Think about and own the second- and third-order effects
How will autonomous vehicles determine what actions to take if they can’t avoid an accident? How will robotics affect jobs and economies? How much autonomy should machines have in decision making?
These questions are being widely debated across many segments of society with industry and interest groups forming to help solve these issues. For example, the Partnership on AI brings together a wide range of experts from industry, non-profit and academic organisations to identify best practices in AI — from ensuring fair, transparent and accountable systems to understanding and addressing potential impacts on labour and society.
Leaders need to be part of these conversations. In addition to participating in industry groups, think about creating an ethics committee that considers the potential uses and misuses of new AI offerings. Encourage and reward critical thinking and debate on AI throughout your organisation. Provide transparency on how data is used and how decisions are made. And back up other leaders when they take a stand on an issue — there have been lots of examples of this recently; we’d love to see more.
8. Embed data security into your culture
Despite well-crafted policies and numerous layers of security, companies are still falling victim to data breaches due to human error. And cybercriminals are only increasing the pressure, using AI tools to escalate attacks. Consequently, leaders must embed data security into the culture. It means ensuring not only that security teams enforce and gamify training to keep employees alert, but also that all leaders proactively manage data risk at the policy, technical, and cultural level as they launch new digital products and services.
Leaders must “walk the walk” on InfoSec, and invest the time and mental energy to be on top of their security agenda every day. One recommendation is for leaders to spend a few days at an InfoSec conference. No-one emerges from these conferences without a 10X increase in paranoia, and only the paranoid survive when it comes to InfoSec.
9. Keep your tech stack up to date
Tools and techniques for data mining, data extraction, data cleansing, and data science that just nine months ago were “cutting edge” are being unseated.
Whether a large company should go with a major vendor or build the DevOps capabilities in house to keep an open source stack up to date will depend on their strategic goals. But, regardless of which direction is chosen, we believe firmly that you can’t fudge the issue.
Investing in proprietary solutions and platforms may mean you’re locked out of new capabilities if your vendors can’t keep up. Equally, “going it alone” with open source will be more challenging from a hiring and capability building point of view, which may mean you move slower.
Either way, challenging your CIO and CTO to make decisions and build a scalable, modular platform for everyday operations as well as experimental machine learning is vital in ensuring your company can remain agile.
10. Always be learning
A common misconception about data is that it offers definitive answers. But really, data doesn’t provide the answer. It only illuminates the path and provides a valuable feedback mechanism for continued learning and improvement. Leaders should ask broad human-centred questions (e.g., How are we working as a team? versus How can we improve productivity?), use data to guide their decisions, and then repeat the process to evaluate outcomes and adapt as needed.
No-one knows what the answer will be before looking at the data. One manufacturing leader we worked with, for example, sought to use advanced analytics to uncover inefficiencies in the manufacturing process. However, the data ultimately revealed that the company’s manufacturing challenges were as much about cultural issues as process.
Being open to the outcomes and using data as a feedback mechanism will enable you to continually improve your organisation’s performance.
The fundamental tenets of leadership — vision, communication, agility, and flexibility — will always be essential. However, it’s dangerous for today’s leaders to consider AI as just another technology. AI brings with it new business opportunities, risks, responsibilities and insights — all of which require a new approach to leadership.
Sam Bourton is Chief Technology Officer of QuantumBlack, Chris Wigley is Chief Operating Officer of QuantumBlack, Simon Williams is Co-Founder and Director of QuantumBlack
