Workday has become the latest in a long line of companies to turn to AI agents for its latest advances.
The company recently unveiled role-based Illuminate Agents now available for Payroll, Contracts Financial Auditing, and Policy, and Agent System of Record (ARS) to manage these all in one place.
TechRadar Pro joined the Workday team in Dublin to hear more.
Problem first
Primarily, Workday is looking to translate its workplace expertise into a service for customers.
“We are in the business of work, so workplace, workforce work, these all fall under work,” says Kathy Pham, Vice President of AI at Workday. “I think why that’s important is as we think about agents and AI we think holistically about everything that makes up the ecosystem.”
“We understand people’s roles, and then we think about where technology is the best fit in different parts of work. I think the biggest takeaway is that we see ourselves as the experts in understanding work across HR, planning, finance, etc, and then we take that understanding we’ve built in the technology.”
AI agents are proving remarkably popular – but businesses still face a range of challenges, especially in managing deployment, ensuring security and compliance, optimising impact, and managing costs.
Without a comprehensive, coherent AI strategy and deployment, organisations will face increased security risks and fragmented operations, undermining the value of AI and wasting its potential.
That’s the issue Workday’s Agent System of Record (ASR) looks to resolve. This one single system can manage AI agents across a company’s workforce, both from Workday and from third-parties, and provide businesses with the “essential tools to help effectively govern, manage, and optimise their digital workforce”.
This includes efficient and secure ways to onboard new AI agents, track their budget, forecast costs, and define responsibilities – ensuring agents are always under control.
“The first thing I always tell folks is to write down what you’re solving for,” Pham says.
“First think about the problems, and then from there say, right, let’s look out. If it’s a big organization, let’s look out at all the different vendors we have and see what technologies or solutions they have. And then from there say, oh, you might have an agent that helps.”
Workday’s AI boasts a 49% increase in financial planning and analysis efficiency, as well as a 52% faster rate to process accounts payable, allowing customers to reclaim time usually spent on mundane tasks or routine actions, and these can be specifically built around a customer’s policies, values, and systems.
The Policy Agent, for example, will digest a company’s corporate policy details and can even preemptively send information to managers and employees depending on what work they are completing, freeing up resources for HR to deal with the more complex enquiries.
“We hear it time and time again for customers, the overhead of answering all those questions, the service desk that has to respond to all of those different queries, that cost is huge,” says Conor McGlynn, a senior director of product management at Workday.
“Our customers have thousands and hundreds of thousands of staff, so Policy Agent in the context of the page in Workday, can answer any of those questions for that user immediately without bothering the service desk or the person beside them. We’re reducing that cost, directly reducing that cost, with fewer tickets raised.”
Human and agent collaboration
Workday’s ASR brings one central command centre to onboarding, role assignment, management, and even off-boarding roles – enabling a firm to leverage role-based agents from one single platform.
“What’s important here is human and computer collaboration. It’s human and agent collaboration. It’s really a digital workforce and how that digital workforce can be managed through the Agent System of Record,” adds McGlynn.
The ASR works alongside Illuminate, speeding up manual tasks and assisting workers, streamlining processes for productivity and efficiency across the board.
“We’re saving customers time, we’re reducing costs, or we’re improving the quality of the outcome. Like I’m hiring, recruiting, shortening that time down for somebody submitting their expenses, making it quicker than ever before, saving people’s time,“ he explains.
This presents a risk, of course. When machines manage humans, a key consideration is accountability – how can you ensure that your AI Agent is making decisions aligned with your company’s values and processes?
“The accountability and responsibility remains with humans,” argues Workday’s Chair of Technology & Society, Taha Yasseri.
“How much alignment you see between the outcome of the machine and what a human or a group of humans would produce. I think our reference point is always humans. But your question is very interesting because when it comes, for example, to discriminatory behavior.”
AI has an inherent tendency towards bias that needs to be closely monitored and controlled, and this reflects existing societal leanings rooted in issues like gender, race, culture, or politics. This is why Workday says human oversight is an absolute necessity for models like Illuminate.
“Humans are not necessarily the best benchmark. We are not super happy with what we [humans] do. But at least we can make sure that the machines are as bad as us, or maybe even better, but definitely not worse. So the accountability and responsibility remains with humans. I don’t have any sort of art of the vice philosophical thinking here.”
Clean, crisp data
One of the key features of Workday’s model is its data set. “We were born in the cloud,” claims Enda Dowling, SVP, Product Engineering.
“All our data has always been in one place and only one place. It’s a complete data set. And arguably we have one of the cleanest data sets on the planet for HR finance.”
This is important, Dowling says, because incomplete or incorrect data translates into agents that get it wrong, and when you’re managing people, especially from an HR perspective, flawed information can be seriously damaging.
“You need that right, clean, crisp data set to train your models, to help your ML models make better decisions, especially in that agentic world.”
With 20 years of data, context, and expertise, Workday has built its AI right into the core of its service, boasting a “unique knowledge around identifying skills and organisations models in the agentic world” that’s reflected in Workday’s reputation.
“And then of course, customers trust us. Again, it’s that proven track record in terms of 95% plus [customer satisfaction rate]. But you know, that’s not important to us, even though it is. What’s important is that we’re delivering solutions, including agent-to-guides to move organizations forever forward.”
You might also like
https://cdn.mos.cms.futurecdn.net/Zfq9hDycNZa2q8iErM82se.jpg
Source link
