Why Digital Workers Can Finally Handle The Boring Stuff So You Can Do The Valuable Work

Forget brittle rules that collapse on the first exception. We dive into agentic AI for HR and L&D and show how goal‑driven digital workers can process CVs, update HRIS records, draft communications, and handle follow‑ups while staying within clear guardrails and audit trails. 

Instead of chasing speed for its own sake, we make the case for redesigning work: orchestrating multiple agents for recruiting, learning operations, and compliance, with a monitoring agent watching everything 24/7 and a human stepping in for judgment calls.

TLDR / At A Glance:

• clear definition of agentic AI for office work
• contrast between RPA rules and goal‑based agents
• HR and L&D use cases across inbox, HRIS and comms
• human‑in‑the‑loop approvals and escalation points
• multi‑agent orchestration with risk and governance oversight
• redesigning processes instead of speeding old workflows
• adoption mindset, access control and data security
• start‑small pilots and simple agent examples
• time savings redirected to strategic HR and L&D work
• upcoming two‑day Belfast event for HR and L&D

Kieran Gilmurray joins us to break down the practical difference between RPA and agentic systems, why resilience matters, and how HR teams can sequence adoption without risking data or trust. 

We talk through real use cases across inbox triage, data entry, onboarding, and scheduling, then show how to layer approvals, role‑based access, and digital footprints that actually improve governance. 

The result is not a “robot takeover,” but a credible way to give people leaders back 10–30 hours a week for strategic work: workforce planning, skills frameworks, leadership pipelines, and culture.

You’ll hear a simple route to get started: map a day‑in‑the‑life, target the 50–80 percent repeatable tasks, build one agent that delivers measurable time savings, and expand to a multi‑agent system once the value is clear. 

With modern point‑and‑click tools, you don’t need to be a developer to build useful assistants; you need clarity on outcomes, sensible safeguards, and a plan to scale what works. 

If you’re ready to earn a stronger seat at the table where people, tech, and business strategy meet, this conversation gives you the playbook. If it resonates, subscribe, share with a colleague, and leave a review telling us the first HR task you’d hand to a digital worker.

Exciting New HI for HR and L&D Professionals Course:

Ready to move beyond theory and develop practical AI skills for your HR or L&D role? We're excited to announce our upcoming two-day workshop specifically designed for HR and L&D professionals who want to confidently lead AI implementation in their organizations. 

Join us in November at the beautiful MCS Group offices in Belfast for hands-on learning that will transform how you approach AI strategy. 

Check here details on how to register for this limited-capacity event - https://kierangilmurray.com/hrevent/ or chat https://calendly.com/kierangilmurray/hrldai-leadership-and-development

Chapters

• 0:00 Setting The Scene And Guest Intro
• 0:29 What Agentic AI Actually Is
• 1:15 From RPA Rules To Goal-Driven Agents
• 2:34 HR Use Cases And Autonomy
• 4:10 Multi‑Agent Systems And Governance
• 5:16 Redesigning Work, Not Just Speeding It Up
• 6:21 Adoption Mindset, Risk And Access
• 7:15 Start Small: Practical Steps For HR
• 11:34 Tools, Skills And Upcoming Event

Transcript

Claire Nutt: 0:03

Welcome back everyone to our podcast. I'm delighted to be joined again by Kieran Gilmurray. Kieran, tell us what country you're in today.

Kieran Gilmurray: 0:12

This time I'm in Riyadh, just back from Abu Dhabi. I'm here for about a week and then I'm off again. But in the meantime, I visited Egypt virtually, Ireland virtually, and a whole host of other countries that I can't even tell you where I've been. I've got over that.

Claire Nutt: 0:26

Educating one man educating the world. I love it. I love it. So today we're back on our podcast for our HR and L D professionals, and we're focusing in on the topic of agentic AI. And I know you've recently released a new book, your third, and this is a hot topic for you and for a lot of organizations at the minute. So what I really want to get to grips with and simplify, I suppose, is what is agencai AI when it comes to HR and LD professionals? So what is that? A few lines that you could help describe it for them.

Kieran Gilmurray: 1:00

Yeah, hopefully I can do it in a few lines, but it's basically a digital worker. So we're used to seeing physical robots sitting in factories. We're all familiar with the Amazon robots or the Teslas robots that wander around doing things that normally humans would have done. Agentic AI is the same except for the office environment. So the tasks that we perform, for example, you know, opening up an email, downloading a CV, transcribing that, putting it into an L and D or HR system, writing an email, writing a letter and a follow-up, all that is a routine set of patterns, dare I say. And agentic labor, digital labor, can do the exact same thing as a person. Slightly different from what people might have been used to a couple of years ago, which was robotic process automation, which was largely the first version of this to some degree. That was very, very rule-based. In other words, if this happens, do that. And if that doesn't happen, it broke. Now this digital workforce is built on the same technology as Chat GPT. In other words, large language models, you give it a goal and it goes and executes the goal. Now it can come do it by itself as long as everything is logical and clear and digital. But if it runs into mistakes, it can come back and check, or you can code these things to say, okay, you can go so far, but I want to see what you've done, then come back to me and get some more direction and then go again until you're happy that the task that was asked for was completed successfully. And whatever way you try and measure it, it was actually done in the right way.

Claire Nutt: 2:32

And I love that. I love how it doesn't just fall over. Um, I'm very familiar with process automation and flows that just die because something's happened or it's lost connection to something along the way. The fact that a genetic AI can now do this autonomously without someone constantly saying, yes, move to the next step. Yes, move to the next step. I think that brings a new layer of excitement to AI, especially for me because I'm I very much have my head in HR and LD processes. That's where I help organisations from a strategic level down to understand what their business does and help them to create efficiencies. This for me is just that little sprinkle on top of that AI cake that's already there. Um, it's really starting to bring together the decision-making process and alleviate some of those mundane repetitive decisions that HR and LD continuously need to make. And you've given some fantastic examples there already: scanning emails, downloading CVs, uploading that information into a HR information system. And when I start to think about HR processes, there's some great systems out there, yeah, that have fantastic workflows or robotic processes already built into them as well. But generally, there's still a human and quite a few touch points throughout that process. So the fact that HR can train this, they can actually train the logic in it as well. They can train it to check it. And you might be able to talk a little bit more about this, Kieran. But from what I can gather as well, we can actually train security agentics to actually keep an eye on everything and make sure it's safe.

Kieran Gilmurray: 4:16

So you don't necessarily have to have just one digital worker that can do everything. So you could have, you know, a HR digital worker or agentic agent, you could have an LD agenc agent, and you could have a governance or risk agent monitoring the work that the other two. So if you ignore the word agent for a moment or digital worker, and imagine how you construct your real world, then all you're getting robots to do are the work that you and I would have done previously. Now, that should excite and not frighten people because let's be honest, nobody wants to sit and do the mundane and repetitive. Now it is necessary work, but it's not why you and I get out of bed in the morning. We don't go yippee, let me do another form download. So if we can, as you say, get this digital labor in doing the work, getting monitored by other agents, and if you wish, putting a human in the loop, in other words, getting a person to approve everything as well, then there's no reason why you wouldn't reconstruct how you build your HR department. And again, other departments in the organization that involve you know routine, mundane, processural digital work. Where any of that exists, you can now put a digital worker accompanying a person or instead of a person to get the job done 24-7. So it's kind of exciting technology. If last year was about LLMs, this year very much is about agentic AI and expect to hear more about that in the coming years as well.

Claire Nutt: 5:40

Yeah, it is very exciting. And I this starts to get me thinking about what HR and LD could see as their potential opportunities, and you've already touched on it exactly there, redesigning the way that we work. And we've said before in previous podcasts, introducing AI is exactly like introducing any form of new technology. You don't just hand it over, you do not give everyone access, you do not give it access to everyone's data, you do not train people. Um and I think organizations sometimes, when it comes to AI, are overthinking this. It's a new piece of technology that will assist the organization, and the exact same principles need to apply when it comes to risk, security, access, and use. And I think if organizations can start to think at from a strategic level, okay, what work do we actually do? What are the conjoined processes that lead to a day in the life of LD? Now they will never give you the exact same answer, it will be different every single day. But if 80% of those tasks were repeated once a week, is that something potentially then that a genetics could be introduced into? I think so. And I think it's just putting that first foot in to say, okay, what do we do? How do we structure our work? And then ripping that up and starting again. If we were to write this, what would it look like? You could have 10 of these agents working, as you say, um, in in on many different processes, but all being monitored and then overseen by a human in that process too. And you and I have said this, you know, AI is about augmenting the human role. And I do believe that agentic AI is the same. I think it it brings the same opportunity. Yes, there are always degrees of risk um with compliance and things like that when it comes to technology, but I still think it brings the same level of opportunity and it should not scare people. It's just thinking about it differently now as well.

Kieran Gilmurray: 7:41

Now you've got though, don't you? You've sort of a uh a basket of goodies here that can do anything, you know what I mean? And I love what you're saying there, which is like don't redesign your processes to do what you're doing to do things faster, because that's somewhat faster rubbish. You know, it gives you a moment to add in digital labor and digital intelligence to do things entirely differently. And it's not new, by the way, as well. We should warn people hearing these words, but if you've been driving a Tesla car for a couple of years, that's essentially a version of agentic labor. You give it a direction and instruction, it goes off and does it. But you're right, like you know, I sometimes think we set too high a bar, Claire, on robots. You know, if it's not 100% right, we won't do it. But when we rephrase that and said, look, if you had a contact center with 100 people and I freed up 80 to do more valuable work, would you love it? Yes, I all day long. Well, if agentic ARAI gets it right 80% of the time, oh no, you know, it has to get it right 100%. You know, I I think I do think we need to view it differently to get the benefit of it, rather than just making things that little bit quicker. But what you were describing there is called a multi-agent system, and multi-agent systems need orchestrate, orchestrate it. The good news is you can get another agent to orchestrate all of the agents, you could have a risk agent actually monitoring what they're doing in real time 24-7, alerting you, whereas you and I are not available to check what everybody is doing in their own team. But because it's all digital, it's all digitally tracked, there's a digitally fit a digital footprint, and no agent can do what it wants by itself. It has to be told what to do. So it's never going to go rogue, it's never going to go down the pub and gossip or hand or leave documents behind or a laptop. So actually, it is an opportunity to put agents into workflows where if they can handle 80% of your work, brilliant. If they can handle 50% of your work, even better still. Because imagine having another 10, 20, 30 hours a week free to do the stuff that you always wanted to do. That's why it's such an exciting opportunity, and the technology really works.

Claire Nutt: 9:40

And that's that is what we hear HR and L and D talk about time. You know, what keeps them up at night, not having enough time to do the things that they want to do in those working days, or they've some employee relations issues over here that are taking them away from this strategic directive? It's it's always about trying to do things better, but also trying to save that time so they can focus on strategic opportunities as well. So, Karen, if you were to wrap up our conversation about agentic AI for HR and L and D, what would you say that HR and LD professionals need to do now really to start to prepare for that agentic labour system?

Kieran Gilmurray: 10:20

Yeah, I think they, you know, the older dash, Claire, you know, think big but start small to build a digital muscle is true, because you can build an agent within ChatGPT or within, you know, Gemini. It's called a gem. Uh, for example, I have an agent every day, every time I have a meeting, the agent does a bit of research on the person, does a bit of research on the topic, outlines what I'm actually going to do for the entire day, makes a suggestion based on review of previous meetings as to how to optimize it and get the outcomes that I want. And all that was pretty much built in an afternoon with a point-click, point-click. That's a very simple example, but it's freed me up a couple of hours a day of research. Now, now I can have the time to reinvest further, where I've looked at all of the work that I do, uh all of the work that I want to do, all of the work that I enjoy, and the work that I don't enjoy either. Because in this instance, everybody goes, What if I hate? I'll just get a robot to do that. That's sometimes it isn't very valuable, doesn't take very long. And if you actually did it, you'd probably get it done in half an hour. But we hate doing it, so it takes about six. But if you then start to look at all the work you're doing and say, where can I reasonably hand that off to if it was a human, fantastic. But this is not a human, by the way. But you've now got a digital assistant where for the first time ever, for not a lot of dollars a month or euro or pounds or whatever your denomination is, you can do that. You have got a really intelligent Jarvis or whatever you want to call it that is doing work for you to free you up to do the things that really matter, all the strategic stuff that the business values. And this is the type of thing that you need to be doing to get a seat at the board to earn HR and L and D's rightful place, which is we need a people strategy, we need a tech strategy, and we need a business strategy. That three-legged stool has to exist. Each part has to play an equal role in the growth of the organization, and agentic labor now frees you to allow you to do that.

Claire Nutt: 12:11

Absolutely. And some of those programs that you're talking about have really nice user interfaces. So we should really say you do not have to be a tech expert to be able to build these either. Um they can be really straightforward. And I'm so excited that we're going to be talking a little bit more about Agentic AI and our upcoming two-day event for HR and LD professionals in Belfast in November. Great to catch up with you, Kieran, and stay tuned for our next podcast for HR and LD professionals.

Kieran Gilmurray: 12:41

Thank you.