Creating the Ultimate AI-HR Tech Stack
Starting fresh.
Welcome to FullStack HR, and an extra welcome to the 302 people who have signed up since last week. If you haven’t yet subscribed, join the 8000+ smart, curious HR folks by subscribing here:
Happy Monday!
I’ve been in a horrible writing slump, and for the first time in three years, I’m taking my biggest “break” from writing since starting FullStack HR.
I aim to double down on writing for the last stretch of 2024. So, what have I been doing instead? Talking, talking, and working, working. It’s truly a blessing to work with HR, meet some of the world's most exciting companies, and help them use AI.
I still pinch myself that I’m part of this—and I’m not done yet!
I’m hosting a free 90-minute prompt BootCamp for HR professionals today!
Invite your team, bring out ChatGPT, and join in!
As I said, it’s free and very practical.
But for now, let’s dive into today’s article: What would our tech stack look like if we had a blank slate?
Let’s get to it.
What would it look like if you started again, building the people department from scratch with AI tools and an AI mindset?
I’ll take a jab at describing what I’d do if I had a blank canvas to create on. (With an almost unlimited budget - a utopia, I know, but still.)
This is more of a thought experiment than anything else, but here’s my take.
Assuming I’m the first one, my starting point would be getting a ChatGPT Enterprise license.
While Claude, Copilot, and Gemini are making strides, ChatGPT Enterprise remains the king of the hill in my book.
I’d use it as the foundation for everything I want to build.
New recruitment process? ChatGPT.
Onboarding plans? ChatGPT.
Compensation and benefits policy? ChatGPT.
And the list goes on.
It’s like the ultimate multi-tool that all the cool kids had back in the day—it gets the job done and is great for many things. Plus, it’s easy and instant. For this reason, ChatGPT would be my starting point. From there, I’d leverage its insights to build a model, most likely turning to ThoughtfulGPT.
On top of that I would also utilize Sana. Why? I’ve written about that in the past.
I would also (and this is not easy for me to say) use Microsoft Office 365 over Google and utilize the Copilot features in MS Office 365.
With those three, ChatGPT, Sana and MS Office 365, I could probably make do for quite a while, combined they are that good utilzied right.
But that doesn’t create a great article, does it? No. So let’s move on.
Hiring
Let’s assume we need to hire people. Hiring has historically been quite expensive. I’d start by defining the recruitment process with the help of ChatGPT. Then, depending on the role, I’d use tools like Holly or TalentRiver to integrate AI early in the funnel to find candidates. For streamlining interviews, I’d turn to Metaview.
ATS? I would utilize something that has AI built in from the get-go, and there are tons of great systems out there. (My Swedish heart beats for TeamTailor.)
Ideally, I’d also use something like phonescreen.ai to vet candidates, but that would depend on the roles we were hiring for.
All these tools for recruitment? What about the cost?
Given how expensive it is to engage third-party recruitment vendors to do similar work, I’d argue these tools pay for themselves if they result in just two hires.
And don’t get me wrong here, I would still have to do the hiring but they would agument me to help me do a lot of the work and a lot of the heavy lifting.
But then what?
A payroll system of some sort, paying people is usually appreciated so that would be the first thing (or maybe even before) hiring our first couple of employees. I’d choose a system that is insanely API-friendly.
Once we’ve used these tools and grown the team, I’d invest in a solution to handle employees’ day-to-day questions.
This would free up my time as the sole HR person and allow me to utilize the policies and processes I created with ChatGPT fully. My tool of choice here would be Harriet.
Investing in something like this upfront saves time more than anything. Having basic questions answered or triaged to me would delay the need to hire additional HR people, improving the P&L. However, for these tools to work effectively, they need data.
That’s why you first need to build all the processes and document your thoughts before buying these systems—which brings us back to ChatGPT Enterprise as the foundation.
I would also utilize Tilda when ideating around brining people onboard, train them etc. It’s a great (and affordable) way of helping me creating learning paths and ways of aquiring knowledge in a fast, and dare I say agile way.
What about an HR system?
Well, that goes back to the payroll system's API friendliness. If that system can integrate with Slack or Teams so people can do time deviations directly there, I wouldn’t buy an HR system.
I would pursue the Klarna path with a good, integrated payroll system (they use Deel) and build stuff on top of that, potentially my own generative AI tool to help me utilize the data we would have in the payroll system, at least to start with.
Is this all? Not at all, but it’s a good start!
(It also points to a separate problem that more organizations are struggling with - maintaining all of these HR-AI systems!)