AI and the Efficient Frontier of Risk

One of the most important contributions leaders can make to further the success of their teams is to focus them on the efficient frontier risk.  Taking too little risk is the bane of publicly held companies, causing them to lag behind or miss catching the next big wave.  Taking too much risk is the bane of start-ups, particularly those that have been over-funded, causing, well, we have all seen the train wrecks.  The goal is to find a Goldilocks zone where the risk pairs well with the probability of success and the expected ROI, and where alpha risk, the risk of taking the action, has been properly weighed against beta risk, the risk of not taking it. 

The challenge in finding this Goldilocks zone is that there is no hard data for determining where it actually resides.  A handful of great entrepreneurs have shown they have a sixth sense for it, but the rest of us have to proceed by trial and error.  We are better served by a “wisdom of crowds” approach, exposing the decision to a broader population through open discussion, working toward an actionable consensus.  Federating risk evaluation helps dilute and diffuse biases at the top, and involving the team in the decision-making sows the seeds for genuine buy-in during the execution phase, even when the choice is not one they themselves would have made. 

When it comes time to execute, the zone management framework offers a clear path to follow.  The Incubation Zone is designed specifically to support the iterative, trial-and-error phase of risk exploration.  This is a fuzzy front end, and it does not come to closure until you have crossed the chasm and found at least one compelling use case where taking the risk has proven to pay off.  That is the beachhead use case.  It is your ticket out of the Incubation Zone.  But that still begs the question, where are you going to go next?

Each of the remaining three zones is designed to address a different level of risk.   Depending on whether you assess the risk as high, medium, or low, the zone most suited to hosting your next step will be the Transformation, Performance, or Productivity Zone, respectively.

• The Transformation Zone is used primarily to address an existential threat to the future of your enterprise, one that requires you to reframe your operating model, and perhaps even your business model, if you maintain your viability.  Here, alpha risk and beta risk are both very high, making it critical for the entire team to unite in prioritizing this initiative above everything else on their plate.  The leader has to take full command here, and everybody else has to row hard—no opting out, no passengers just along for the ride.
• The Performance Zone is where the majority of consequential bets get made and where leaders can show their true stuff.  The playbook calls for committing to medium-risk bets (not moonshots) and winning the majority of them.  We normally call this good execution, but we should note it is of a special kind, one that calls for reengineering an established process, redirecting an established team, and keeping both on course until the new normal sets in.  Course-correcting along the way is often required.  Hesitating after things have been kicked off, on the other hand, is a mistake, and typically not one that can be recovered from.
• The Productivity Zone is where the largest number of AI bets are being made today, but I have to say I am uneasy about both the why and the how.  On the why front, I worry that established players are shying away from higher risk bets simply because their cultures have become Productivity Zone first, defaulting to a risk-averse mindset.  On the how front, I worry that the bulk of the AI bets getting made are focused on cost-reducing an existing legacy process instead of redesigning it to eliminate friction, reduce response time, and deliver higher value to the customer.  Setting such concerns aside, however, there is a lot of low-hanging fruit in our back-office processes that is ripe for plucking, and Productivity Zone leaders should certainly be harvesting some of it this year.

If these ideas strike a chord, let me suggest some follow-up steps to take:

1. Take an inventory of all the AI initiatives you currently have underway or under consideration.
2. Assign each initiative to the zone that best matches its efficient frontier of risk.
3. Prioritize your commitments and resource allocation based on the level of risk, working through the higher risk options first.
4. For every initiative that is a go, assign it to a single accountable leader who works in the assigned zone.
5. Assign governance responsibility to an AI change management team, with a “win or learn” charter, and convene it twice a month for the foreseeable future.

One final point: Act now!  2026 is the year of AI, like it or not, so sitting things out is simply not an option.

That’s what I think.  What do you think?