Organizational (re)design choices by tech companies are changing the role of software engineering leader in ways that make the job harder--and generative AI won't fill the gaps. Flatter organizations deprioritize developing people, which ultimately harms project delivery and undermines strategy as well. This piece by Meg Adams on the necessity of a shopkeeper mentality for software engineering leaders not only provides an effective leadership model but a description of the things we can do in the role which generative AI can't.  

We need to do what we actually say our job is instead of spending our time in ways an executive without proximity to the work might think generative AI can replace.  "Scanning” our work domain, adapting to the reality of today’s hybrid and distributed workplaces is how we can build, refine, and maintain the context necessary to deliver projects, drive strategy, and develop people.  People can delegate—but maintain ownership and accountability. AI can't. We are uniquely capable of scanning--but also of micromanaging.

Flatter organizations can't know what they need to about how their people are doing, or where the friction is, or what they haven’t been seeing because they’ve eliminated the managers who could tell them and they lack the systems or signals that could tell them. Friction in key communication channels within the organization and blockers to work progress may eventually show up in other ways, but perhaps not before missed deadlines or costly mitigations occur.  Why?  Because the remaining middle managers in a flattened organizational structure have too wide a span of control to proactively identify and address issues that pose obstacles to work progress. In addition to the visibility issues, flatter organizations eliminate growth paths for employees. What does professional growth look like in an organization where you’re either an individual contributor or the CEO with little or nothing in between except generative AI? Where will the next senior leaders come from?

Believing generative AI will provide credible answers to these key questions seems exceedingly optimistic.  What we’ve seen so far of how the efforts by companies trying to replace entry-level workers with AI does not inspire confidence.  Senior engineers in these companies are burning out from overwork and leaving to find employers that enable them to better balance their personal and professional lives.  Senior engineering leaders who face the prospect of directly managing dozens of people may make similar decisions rather than suffer burnout. Companies betting that generative AI can fill the gap are risking their long-term viability.