Companies of all sizes are talking about using AI to flatten their organizations—and eliminating an entire layer of middle management in the process. Gartner projects that 20% of companies will cut more than half of their middle managers by the end of the year. 

But that might be throwing the baby out with the bathwater. A far better approach is to keep middle managers as the important link between leadership and frontline employees that they’ve been for decades. The key is to also ensure they have the AI skills they need to drive better business outcomes in this new AI-enabled era.

From Doers to Orchestrators

So, what do they need to be trained on to meet the moment? They’re no longer the hands-on coders or “doers” of the organization; they’re the orchestrators that make sure business objectives are being met. Where might they now need to augment their skills?

First and foremost, they need to be trained on what AI can and cannot do. Middle managers cannot tell their teams to “use AI to work smarter” if they themselves don’t have a baseline understanding of what AI tools currently exist and how they can be deployed within their particular industry. 

For example, in the finance department, AI might be useful for simplifying tax return preparation but not as effective right off the bat in complex audit scenarios. Or on the manufacturing side, it might be useful in predicting maintenance needs but not inspections.

The point is, any business or industry has a lot of nuance, and it’s important to understand where AI can fit in, or even whether there is a need for it.

Clarity on Business Objectives

A second important skill is having the business acumen to identify the right use cases for AI within a business line, department or division. What are you actually trying to achieve? Revenue growth? Cost reductions? Productivity enhancements? Customer retention? Faster time to market? Middle managers need to be totally clear on their business objective before they even bring AI into the picture to automate a process or work towards a particular goal.

Once a team is using AI for a well-defined, objective-aligned use case, middle managers will need to develop a new supervisory skill: overseeing their people’s ability to generate accurate, high-quality AI outputs.

A key part of this supervision is making it clear that employees remain fully responsible for quality control and for verifying the accuracy of any AI-generated output. Falling back on “This is what the AI gave me” cannot be an acceptable response. That message has to come through loud and clear so that there’s an embedded culture of accountability around AI usage.

Similarly, to effectively coach their teams towards quality outputs, middle managers themselves need solid skills in prompt engineering—knowing how to set the right parameters and refine inputs to get the best results.

That starts with knowing what question you actually want answered. For instance, pointing AI at a spreadsheet of sales prospects and asking it, “Who are my best target customers?” is too vague a question. Far better to ask AI, “Who are my best target customers within Geo X who have annual revenues of at least X amount and have also engaged with our website within the past X months?” Iterating the prompt and adjusting the parameters produces a much more specific and actionable output.

People and Processes

Ultimately, this new era demands that middle managers become skilled not just at managing people, but at managing processes. They need to become better workflow designers: looking at KPIs, looking at AI, and then figuring out how to bring the two together. From there, it’s a matter of execution by the team, with middle management guiding and coaching them along the way to ensure they’re effectively leveraging AI to obtain the desired results.

This evolution of the middle management role is needed more than ever, because the gap between what organizations need and what they’re currently achieving is already visible. A recent research report surveying 150 C-Suite leaders found that many organizations are struggling to translate their investment in AI into meaningful business impact.

Part of the reason, according to the report, is that embedding AI at scale requires both job redesign and a cultural adaptation as much as a technical implementation – something that executives can mandate from above, but middle management is ideally positioned to carry out.

When middle managers have the skills to understand the technology, to redesign workflows to take advantage of it, and to coach their teams toward higher-quality outputs, AI stops being a question mark and becomes a powerful strategic asset.

At that point, the conversation shifts from “What are we going to do about AI?” to “What are we going to achieve with AI?”—and that’s exactly where today’s companies need to be.