Agentic AI is already old news: Why Generalist agents will transform HR
Say goodbye to siloed tools—meet the next generation of AI colleagues.
Welcome to FullStack HR, and an extra welcome to the 172 people who have signed up since last week. If you haven’t yet subscribed, join the 8100+ smart, curious HR folks by subscribing here:
Happy Thursday!
Christmas is heading towards us, as per usual with the speed of an express train.
I’ll do one last webinar before Christmas.
Free as always, focused on AI and HR, as always - join in on LinkedIn on the 19th of Dec - 3 PM CET / 2 PM GMT / 9 AM EST.
(or, to put it another way, 15.00 Swedish time, just after tea time in the UK and basically in the morning for people in the US.)
But I also have a small favor to ask today—maybe a day late, but let’s give it a shot anyway.
If you’ve read my newsletter, joined one of my webinars, or listened to me speak live (IRL, as we used to say) and truly enjoyed it, I’d be incredibly grateful if you could head over to the Transform Awards site and nominate me for this category:
It would mean a lot if just one or two of you could take a moment to drop in a nomination.
(Yes, it feels a bit awkward to ask, but I’m challenging my Swedish humility here!)
Thanks so much for considering it!
Now, let’s dive into today’s article.
If you’ve been paying attention, the chatter this fall has been about AI agents. Josh Bersin has sung their praises, and the major players—SAP, Workday, you name it—are rolling out AI-powered assistants designed to make our HR lives easier.
The narrative? These are highly specialized tools embedded in HR platforms, tailored to solve specific tasks with the precision of a Swiss watch. But as we head toward 2025, I think we might be asking the wrong question.
It’s not about how many niche AI agents we can stuff into our systems—it’s about whether we need them at all. What if the future of AI in HR isn’t about more specialized agents but fewer, smarter ones?
Specialist AI agents
Okay, so let’s say you are a Workday customer. You’re in Workday, trying to get a report on attrition trends. Instead of digging through menus, an AI agent pops up and does it for you. It’s perfectly integrated into your system, knows your data, and promises to save you hours.
Sounds great, right? That’s the vision these specialist agents offer. They’re like highly trained assistants who live inside your HR software and know your workflows inside and out. As said, companies are already rolling out these features, and they’re a natural evolution of how we think about automation.
But here’s the catch—these tools are only as good as the system they’re built into. What happens when you need to go beyond Workday? When you want something that works across platforms, connecting the dots between your HRIS, your ATS, and whatever else is in your tech stack? And I think a very likely scenario is that you have multiple systems from different vendors (even though the major players want to go all-in Sauron and have one system to rule them all).
Generalist AI agents
Enter generalist AI agents. These aren’t tied to a single system; they’re built to work like a jack-of-all-trades, jumping between tools and tasks like a human would.
What has sparked this train of thought for me is Claude’s Computer Use. I’ve talked briefly about it in the past, but if you haven’t laid your eyes on what it is yet, watch this video of it:
And it’s not perfect (ask me about its crashes - I was about to throw my computer out of the window at one point), but it shows what’s possible. Say you’re scheduling interviews across Teams, Google Calendar, and some random third-party app. A general AI agent could:
Sync all the calendars
Juggle everyone’s schedules
Book the interviews
Handle follow-ups
No need to manually copy-paste or double-check—it just gets it done. And that’s just one example.
This reminds me of the debate around language models. Sure, specialized models are great at one thing, but general-purpose models like GPT-4 can handle a lot more if you know how to use them. The key? How you interact with them
Similarly, a generalist AI agent might not be built for just one system, but its ability to adapt and connect makes it way more versatile.
So, where are we headed?
I see two paths:
Specialized Tools Dominate (Short-Term)
Big platforms like Workday and SAP will refine their AI assistants.
They’ll excel at system-specific tasks.
HR teams will see value in their immediate ROI.
Generalist Agents Take Over (Long-Term)
These will be the Swiss Army knives of AI—able to do it all.
They’ll operate across platforms, like a colleague who knows all the systems.
The combination of adaptability and expertise will make them indispensable.
What excites me most is that generalist AI agents reflect how humans work. We don’t limit ourselves to one tool or task. We switch between systems, connect dots, and adapt on the fly.
Imagine an AI agent that isn’t just a tool but a partner—someone (or something?) that:
Understands your company’s culture and context.
Bridges gaps between HR, finance, and IT (or essentially all departments).
Learns and evolves as you do.
That’s the kind of future I want to see in HR tech. Add a proprietary model on top of it all and we have an winning strategy. (And yes, I know nothing of this is easy but still - it’s the future I would want to build towards.)
And no, specialist AI agents aren’t going anywhere—they’re incredibly useful for now. But I think they’re just a step toward something bigger. Generalist AI agents might feel like a far-off dream, but they’re closer than you think.
If you’re leading an HR team, it’s time to start preparing. Think beyond tools tied to specific systems and imagine a world where AI works alongside you, not just for you.
Because let’s be real—the last thing we need is more disconnected systems. What we need are smarter, more adaptable AI colleagues who make our jobs (and lives) easier.
We let the general AI agent do the grunt work, and we’ll focus our efforts on more fun and more business driving initiatives.