Eight steps to enterprise AI leadership

Photo credit: Ryan Loughlin via Unsplash

True AI expertise is rare. AI leadership expertise, the ability to lead organisations which have embedded AI into their operating models, is even rarer. And enterprise AI leadership, the ability to lead the adoption and use of AI within established businesses, is so rare as to be virtually non-existent.

This rarity seems to persist, despite the number of press releases, articles and social media posts which claim that ‘AI is transforming industries’. My perception is that most leaders in most industries would admit that they find AI as baffling as it is exciting, as frightening as it is promising, and that they wish that it would slow down for long enough for them to catch up.

I believe that it is possible for enterprise leaders to catch up with AI (whether it slows down or not), but that it requires as much change in their skills, behaviours and attitudes as it does in their technology, processes and organisations. Enterprise leaders will not become enterprise AI leaders unless they change themselves as much as their businesses.

Many of these changes are not new: they are management and leadership practices which have been established for years. But the arrival of AI amplifies the need for these practices.

The change involves eight steps. I will summarise them here, and explain them in more detail in the next eight weeks. I had the privilege of sharing these thoughts with Ajit Jaokar's class at University of Oxford earlier this week, and Ajit and I plan to make them the core of a course to be run at Oxford later this year.

Step one: learning the fundamentals

It might seem that, in the age of AI, you don’t need to understand technology any more: some have suggested that everyone’s a developer, and that no-one needs to learn software engineering. But if you are proposing to give your technology autonomy and agency, if you are going to keep track of a rapidly developing market, and if you are going to let your whole organisation build their own solutions, you had better have a grounding in the basics: enterprise AI leaders need to learn how the machines that run their businesses actually work.

Step two: changing your behaviour

Despite years of talk of empowerment, team building and servant leadership, when people take on leadership roles, they are expected to be confident, certain and in charge. But it’s hard to be any of those things when your business is being reshaped by forces that change every few months. Enterprise AI leaders must learn to listen, adapt and be open about their uncertainty, to combine their confidence and certainty with curiosity and humility.

Step three: building the team

Leaders never lead alone: their first job is to build a team. But, too often, teams are unbalanced. Sometimes leaders treat part of their team as ‘core business’ and other parts of their team - technology, HR and, possibly, AI - as ‘support functions’. Sometimes they delegate the understanding of these topics to these team members, so that they don’t have to understand them themselves. Sometimes they do both at the same time. Enterprise AI leaders must learn to build inclusive, balanced teams, where each member helps the others to understand their specialism.

Step four: tackling the market

Unless you are going to invest billions in building your own cloud, training your own foundational models, or making your own chips, your enterprise will depend on third parties. But the market does not offer the stability that enterprises have traditionally sought: it is hard to know which partners to choose, or whether they will even exist in twelve months' time. Enterprise AI leaders need to learn what drives the market, and how they can make equitable, trustworthy agreements.

Step five: imagining the enterprise

Enterprise ‘transformation’ initiatives usually fail. Too often, they are modest improvements on today, which deliver incremental improvements, but not transformation. Or they are broad-based attempts to do everything everywhere, all at once, through waves of innovation and experimentation that don’t stick. Enterprise AI leaders must learn to look at their enterprises forensically and imaginatively, and identify those areas where AI might make a difference - and what to do if it doesn’t.

Step six: delivering the vision

Vision is hard: delivery is harder. Today, there are more enterprises with stalled AI initiatives than enterprises truly deploying AI at scale. This is a failure of leadership as much as a failure of execution. Enterprise AI leaders must learn to create teams and environments capable of sustaining change over time, and to support them through setbacks as well as success. They must also learn how to lead teams which will continuously disrupt and rediscover their own work.

Step seven: leading with integrity

The bad news for anyone waiting for AI regulation and legislation to tell them how to use it is that it will take longer than they expect, and it will move too slowly to adapt to new developments. The good news is that enterprise AI leaders don’t need to wait for regulation and legislation: they can learn to be clear and explicit about the values which drive them and their enterprises, and transparent about how they ensure that their use of AI follows these values.

Step eight: living with uncertainty.

After all this learning, grasping fundamentals, changing behaviour, building a team, getting to grips with the market, re-imagining their enterprise, delivering what they have imagined and declaring and demonstrating their values, can the enterprise AI leader take a break from learning, and focus on running their enterprise? Unfortunately not: the world will continue to change in unpredictable ways, and the enterprise AI leader needs to learn to live with this volatility.

Taking each of these steps is difficult; taking all eight of them will, for many of us, mean changes to style, skills, culture and self-perception. But taking these steps is better than pretending that nothing is changing. As mentioned above, we will be launching a course on this topic later this year to explore these steps further, and to help people take them: further details to come.