No one can ignore AI, and Workday is no exception. The company does paint an interesting, well-thought-out, and not too hype-sensitive picture of the technology. AI agents may not really be your new colleagues, but it takes a professor of technology and sociology to define the relationship between humans and machines. In the company’s EU hub, bright minds are hard at work contemplating the role of AI in work and society.
Who has never drafted an email with the help of AI? And who has never received an email augmented by AI? “In fact, such an email conversation is not just between two people, but between four actors,” argues professor taha Yasseri. He is associated with both Trinity College in Dublin and the Technological University in the Irish capital, and heads the joint center for the sociology of humans and machines (SOHAM).
Prof. Yasseri is a guest at Workday’s EMEA headquarters in Dublin, where he is one of many speakers who, over a period of two days, tries to clarify the role of AI within Workday and the world. The strong European roots of the HR software specialist are also discussed.
Ten years of AI experience
Enda Dowling, senior vice president for product engineering, emphasizes that Workday was building AI before it was cool. “We’ve been building AI solutions for more than ten years and are very comfortable with it. Moreover, Workday was born in the cloud, so all data has been in one place from the beginning. We have one of the cleanest HR datasets in the world.”
Workday’s AI innovations exist under the brand Workday Illuminate. Think of it a bit like Einstein from Salesforce: both brands refer to a set of AI -related solutions.
“Due to our experience with AI , customer data is already well isolated,” continues Conor McGlynn, senior director for AI UI platform. “We have private and domain-specific AI models, and ensure that AI doesn’t accidentally expose data.”
“It’s important not to develop a solution for which you then have to find a problem,” McGlynn adds. AI should be a technology that solves concrete problems and supports goals. “For example, AI can generate a job description for a vacancy in a few minutes, where it may take several hours for humans to do the same.”
It’s important not to develop a solution for which you then have to find a problem.
Conor McGlynn, senior director AI & ui platform Workday
Ai within the Workday platform can observe what you’re doing. It’s aware of the context and understands the content of screens and forms. This context enables broader assistance and functionality. Workday is working on its own agents but is also partnering with third parties via an API. Even agent-to-agent communication is on the agenda.
An agent colleague?
What exactly an agent is, that’s up for discussion. The Workday experts lament the lack of a clear definition. AI and definitions: it remains a complex task. They do agree that an agent should be able to take action.
“In this way, they should help us to be better at our jobs,” says Kathy Pham, VP for AI . “Some things machines can really do better. We can’t stay awake for 24 hours.”
This is how people and machines are increasingly working hand in hand. “An agent almost has a personality,” says Caroline O’Reilly, VP for agent development. “It can choose and execute its own plan based on context with the tools at its disposal.”
An agent almost has a personality.
Caroline O’Reilly, VP Agent Development Workday
In this context, agents are almost part of the workforce. “Agents are not equal to humans, but they do take on tasks and roles that humans have.”
Organic inclusion
Because of that, Workday has recently introduced a management capability under the name Agent System of Record. “Today you already manage the performance of human talent,” clarifies Clare Hickie, CTO EMEA, “and now you can also manage your digital workforce. Humans and computers will work together.”
In that respect, the integration of agents within the Workday suite is a logical extension, although everyone emphasizes that it is not the ambition to view agents as digital alternatives to humans. AI agents do have synergies with human employees: like human personnel, they must have specific access rights for certain data, while other information should not be accessible. Workday works with payroll data, so security and privacy are sensitive issues.
Sociology with Algorithms
This is where Professor Yasseri comes in again. He sees a new form of sociology emerging. ‘The way agents interact with us has an impact.’ Yasseri conducts part of his research for Workday, where he was hired with the help of a written motivation that was partly optimized by AI, responding to a job description which was also given an AI touch.
The impact on society will be unpredictable.
Professor Taha Yasseri, SOHAM
“The impact on society will be unpredictable,” he continues. “People and machines are already working together today, both in teams and in hierarchical settings, and sometimes even competitively.”
“Companies think about this in the short term, but scientific research moves slower and can think further ahead. The habits we develop in our relationships with machines may impact how we will interact with each other.” Yasseri tackles the problem with a multidisciplinary team of researchers, including physicists, network specialists, and philosophers.
Be Silent About AI
However the relationship between humans and algorithms may unfold, at Workday they hope that the AI hype itself will subside a bit. Pham: “I hope that in five years we won’t be talking about AI any more. People no longer ask about Machine Learning, and companies no longer explicitly position themselves as cloud companies. The technology is simply there, and only comes up in targeted discussions.”
Bernie Foley, VP and product manager for all talent-related software, shares that opinion. “We need to ask the right questions to companies and really understand their problems. Sometimes the right solution can indeed be AI-related, but not everything has to be about that hype. A tweak in a product can also be a solution, or an extension of functionality, whether or not in collaboration with a partner.”
“From the beginning, we think carefully about what we’re building, what we want to achieve, and how,” she continues. “Many elements play a role in this, including sustainability.”
European Stamp
All these matters are co-directed from Workday’s EMEA headquarters in Dublin. That’s the other common thread of the event. Dublin is not only an important research hub (where a large new office is currently being built), but a place where Workday’s heart beats.
read also
Workday launches platform to aggregate AI agents
“I am the VP for Talent,” Foley emphasizes. “And I report directly to the General Manager. I’m involved in making global decisions. On the floor above us, engineers are building an agent for recruiters, and another coach agent. Workday is a global company, but here we influence the global perspective and develop important products.”
More Sincere Discussion
Agents, whether developed in Europe or not, will certainly increase in importance. This applies not only within the Workday portfolio. The agent hype is in full swing and every organization wants to position itself as a specialist in Agentic AI. At Workday, we hear for the first time that there’s consideration of the societal impact of algorithm colleagues, and in a well-founded, scientifically driven manner.
Although Workday talks about AI for two days straight, we notice less inflated hype than with other companies. Workday wants to position itself as an expert in many tools including AI, with European roots in governance and development, and succeeds well in doing so.
Thank God for the Rain
McGlynn knows why that is. He points to the window of the meeting room, where the rain is relentlessly splashing against it. “The bad weather is one of the things that makes us good at AI,” says the Irishman. “It’s our superpower, giving us built-in pessimism and cynicism.”
“If you’re in a beautiful place, by the sea with nice weather, there’s a lot of optimism, but also a lot of hype. That’s all very nice, but a bit of cynicism about the application of AI is needed in that context as well.”
“Making wild statements about the future of AI is not wise,” he concludes. “The conversation needs a bit of bad weather.”
In this context, we should not get carried away. AI algorithms are not real colleagues. ‘We do attribute human characteristics to AI and anthropomorphize it,’ says Prof. Yasseri, ‘But no one is going to claim that machines today are responsible for their decisions. That responsibility remains with humans.’