Auchenglen Consulting

Your finance team just deployed a GenAI tool that automates month-end reporting. Brilliant. But six months later, you’ve still got the same headcount, the same team structure, and the same spans of control. Sound familiar?

This is the conversation I’m having with CHROs and Heads of HR Strategy across multiple sectors. The technology lands, productivity theoretically improves, but the organisational structure remains frozen in place. We’re automating work without reimagining how work gets done.

The Old Playbook Doesn’t Work

Traditional capacity planning followed a straightforward logic: forecast demand, calculate required FTEs, determine optimal spans of control (usually somewhere between 5-8 direct reports), hire accordingly. The variables were predictable because the nature of work was relatively stable.

GenAI has broken this model. When a tool can draft policy documents, analyse workforce data, or respond to routine employee queries, you’re not just making existing roles more efficient. You’re fundamentally changing what those roles entail. A business partner who spent 40% of their time on data analysis and reporting now spends that time differently. But doing what, exactly?

Three Methods for Sizing Teams in an AI-Augmented World

Start with workflow mapping, not headcount

Most organisations approach this backwards. They ask: “How many people do we need?” The better question is: “What work actually needs to be done by humans?” Map your team’s workflows in granular detail. Which tasks require judgement, relationship-building, or novel problem-solving? Which are routine, rules-based, or data-intensive? GenAI excels at the latter, struggles with the former.

One professional services firm I worked with discovered that their HR advisors were spending nearly half their time on tasks that GenAI could handle competently. But rather than cutting headcount, they redirected that capacity towards proactive employee relations work that had been perpetually deprioritised.

Rethink seniority ratios

The traditional pyramid structure assumes that junior staff handle routine work whilst senior staff tackle complex issues. When GenAI absorbs much of that routine work, you need fewer junior roles and different senior roles. I’m seeing organisations shift from 60/30/10 (junior/mid/senior) ratios towards something closer to 40/35/25.

This isn’t just about promotion rates. It’s about fundamentally reconsidering what “junior” means when the learning curve for certain technical tasks has been compressed or eliminated entirely. Do you still need two years of doing routine analytics before someone can contribute strategically?

Build in iteration cycles

Here’s the uncomfortable truth: you won’t get this right the first time. The pace of AI capability development means your carefully calculated capacity plan may be obsolete in 18 months. Rather than treating organisational design as a periodic restructure, treat it as an ongoing experiment.

Set explicit review points. Quarterly is often appropriate for teams heavily impacted by automation. Ask: What’s working? What assumptions have proven wrong? Where are bottlenecks emerging?

The Common Pitfalls

The biggest mistake I see is treating this purely as an efficiency exercise. You implement AI, reduce headcount, call it transformation. Then you discover that the relationship-intensive work that drives actual business value has been neglected, and your employee engagement scores have fallen off a cliff.

Another trap: assuming that spans of control can expand indefinitely because “AI will help managers manage more people.” It won’t. GenAI can help with performance data analysis or scheduling, but it can’t replace the judgement required to develop talent or navigate complex interpersonal dynamics. I’m sceptical of any capacity plan that assumes managers can effectively lead 15+ direct reports simply because they have better tools.

Finally, many organisations underestimate the change management required. Your team structure isn’t just an org chart. It’s embedded in how people understand their careers, their value, and their relationships. Changing it requires genuine engagement, not just a town hall announcement.

The Opportunity Hidden in Plain Sight

The most significant shift I’m seeing is that organisational design has become a genuinely multidisciplinary endeavour. You can’t do this well with HR alone in a room with PowerPoint.

You need technology leaders who understand AI capabilities and limitations. You need finance partners who can model different scenarios and productivity assumptions. You need operations managers who know where the actual bottlenecks are. And you need employees themselves, who often have the clearest view of which tasks are automatable and which require human judgement.

This represents a real opportunity for strategic HR leaders. The question “How should we structure our organisation?” has moved from a periodic strategic consideration to an ongoing operational reality. That means a seat at the table for every significant business decision.

Where to Start

If you’re trying to right-size in the age of automation, begin with a pilot. Choose one team where AI adoption is already underway. Map their workflows in detail. Model different structural scenarios. Test assumptions with small changes before committing to wholesale restructure.

Treat your first attempt as a prototype. You’re not looking for the perfect answer but rather building organisational muscle for ongoing adaptation. The organisations that succeed won’t be those who design the perfect structure today, but those who build the capability to keep redesigning as technology and business needs evolve.

The goal isn’t to do more with less. It’s to do different things, better. That requires rethinking not just how many people you need, but what you need them to do.