It seems like just yesterday that using AI in TA was innovative.
But things have moved fast — really fast — up to the point that it’s more common than not to see an organization automating some area of their hiring process. Which, sure, is partly due to the proven results that AI delivers. But it’s also due to the progress we’ve made to develop that technology.
Thanks to a series of hyper-fast advancements in the development of large language models (LLMs), we’re now surrounded by a saturated market of extremely powerful AI solutions: ChatGPT. Gemini. Claude. Grok. And now the newest revelation, “agentic AI.” You might’ve heard this buzzword going around, but what does it even mean? And more importantly, how can TA leverage it to make an impact in the hiring process?
Let’s dive in.
What is agentic AI and how does it differ from other forms of AI?
In typical buzzword-y fashion, the definition of agentic AI differs based on who you ask. But most people would define it as an LLM-powered solution that can make autonomous, goal-based decisions with minimal to no human interaction.
Now some of you might be wondering, isn’t that just regular, old AI? Sort of.
Most people would define agentic AI as as an LLM-powered solution that can make autonomous, goal-based decisions with minimal to no human interaction.
For the sake of clarity (and this is going to drastically simplify things so keep that in mind) we can group “AI” into three buckets: defined automation systems, LLM-powered non-agentic automation, and agentic automation.
- Defined automation systems follow explicit, predetermined workflows. If X happens, then do Y. Because the automation is under such tight constraints, it can be difficult for these systems to handle ambiguity. Think of a frustrating chatbot experience — if you don’t ask a question that directly correlates with one of the AI’s pathways, it doesn’t know how to respond.
- In contrast, LLM-powered solutions can handle that ambiguity. Because their pathways are less defined and their information sources are more expansive, the AI is able to adapt its responses based on previous context or knowledge.
- Where agentic AI differs from other LLM-powered solutions like ChatGPT is its ability to take action. For example, if you asked an agentic AI to book your flight to Seattle, it could theoretically discern the best possible flight paths to SeaTac (ticket costs and airlines), and make a purchase to book your flight at a time when your calendar is free. Because the AI understands its goal is to book a flight, it will perform the necessary tasks to do that, often without any human interaction.
If you asked ChatGPT the same prompt, it could scour the internet and make some great recommendations, but it couldn’t act as you and book the flight. So there are a few more limitations. Granted, being an amazing information library is still really powerful, and has a lot of positive implications in TA and otherwise.
Why does this matter for TA?
With its ability to take action and complete end-to-end tasks autonomously, agentic AI gives TA teams the opportunity to spend more time doing valuable work.
Imagine that. All the administrative work burdening your plate, completed effortlessly behind-the-scenes. At Paradox, we see a future where agentic AI augments certain tasks — like pulling analytical insights or sending onboarding paperwork — while still giving way for humans to handle more important work: the decisions.
And the funny thing is that this isn’t even future state; Paradox has actually been delivering this type of experience since 2016 with automated interview scheduling. Our AI assistants can determine if a candidate meets minimum criteria and move to schedule them for an interview based on recruiter availability. Without the need to play calendar Tetris, recruiters and hiring managers can spend more time actually talking to candidates.
Of course, LLM development is leveling agentic AI up to entirely new stratospheres of complexity. And the crazy thing is… we don’t know where it’ll end.
Our CEO Adam Godson likes to say that we’re in the “era of chaos” when it comes to AI. And, yeah — I mean, look around. Things are moving so fast that I wouldn’t be surprised if agentic AI got labeled as a new buzzword in three months. There’s just so much uncertainty out there surrounding how far this thing will go.
At least we know one thing for sure: AI will continue to make a difference for talent teams everywhere.
