Workday takes on AI’s human in the room

One of the biggest questions in today’s artificial intelligence debate is whether AI helps workers, or burdens them with rework, burnout, and a loss of original thought.

According to Clare Hickie, Workday CTO for Europe, Middle East and Africa (EMEA), the elephant in the room is the human reality of the shift driven by AI capabilities such as Workday’s. For the global workforce software company, AI creates value when organisations redesign work around technology and strengthen distinct human skills such as judgement, curiosity, and critical thinking.

“Workday’s Beyond Productivity shows a hidden drag on the industry,” she said during a press conference in Dublin last week. “Roughly 37% of the time saved through AI is currently being offset by rework. Employees are spending massive amounts of time correcting poor AI outputs.”

That “rework tax” highlights why AI is not yet delivering at scale, and what organisations must change to unlock value.

Greta Stahl, Workday VP for organisational learning, said: “The challenge that a lot of people are running into is that they’re treating AI as just another software tool to deploy in the organisation, as opposed to thinking about it as something that can cause us to fundamentally shift the way that we’re working, or redesign roles to do work differently.”

Stahl pointed to a structural issue rather than a technical one, arguing that simply adding AI to existing workflows limits the impact.

“If you just bolt AI on as just another tool and a process you already have, it’s not going to deliver the value that you want. It’s something that you have to architect for and plan for to get the most value out of tonight.”

Within Workday, that shift has meant redesigning roles and embedding AI into day-to-day work.

“The upside of that for our employees has been that we’ve shifted the skills that they’re focused on, and they’ve been able to spend more of their time working with stakeholders, doing business consultation, developing business strategy, really applying human judgment.”

From an engineering perspective, the limitations of current AI tools can contribute to the problem.

Graham Abell, Workday VP for software engineering and Ireland site lead, said: “Generic generative AI, which we’re all kind of used to, even in our day to day, gets you a starting point, but it does take an awful lot of rework to get it to what you need.”

The reliance on human judgement becomes more apparent when looking at how AI is used in practice.

Abell gave an example from the rollout of copilots, where internal data showed stronger adoption among senior developers than junior staff. He said the difference came down to experience, as senior developers could quickly assess whether outputs were correct, while junior developers were less confident in judging quality and often reverted to doing the work themselves. The pattern, he said, highlighted the importance of critical thinking and instinct in enabling employees to trust and consistently use AI tools.

That distinction highlights the core theme of the discussion: AI does not replace human capability, but amplifies the need for it. For leadership teams, the shift changes how AI success should be measured.

Stahl said: “We love to talk about time saved, because it’s easy to measure. That makes it a very appealing metric to hang on to, but it begs the question of what are you doing with that time?”

She pointed to engagement, retention, and how employees reinvest time into higher-value work as more meaningful indicators of success.

“What humans bring to the table is the judgment that we apply throughout the process, how we iterate to make sure that that specific outcome solves the problem that we need.”

The human edge

The human advantage becomes more critical as AI systems move closer to decision-making rather than simple task execution.

For Kathy Pham, Workday VP of AI, the focus should not be on replacing human input, but on understanding how people interact with AI in different scenarios.

She told Gadget: “In this [one] scenario, we’ll use AI systems to do something, and you don’t need to have an extra check. And then in this [other] scenario, we’ll use AI system to do something, and you should use it as a complement to your decision making.”

That distinction shifts responsibility from the system to the individual, requiring employees to apply judgement based on context rather than relying on outputs by default.

“There are areas that are deeply human that you want someone to still take the time to review it. It’s a training problem too. You teach your employees when they should do something as a supplement versus when they can click go, and that’s it.”

Pham said the risk can be present in how AI is used, particularly when it replaces the process of learning rather than supporting it.

At the same time, she said AI can expand access to skills and creativity, allowing more people to engage with complex tasks and ideas.

“There is a potential where if you only use AI to constantly leapfrog over the process of learning, you can miss a ton of things. But it definitely doesn’t have to be that way,” she told Gadget.

* Jason Bannier is a data analyst at World Wide Worx and deputy editor of Gadget.co.za. Follow him on Bluesky at @jas2bann.