How do you rebuild HR for an AI-powered future?In this episode of The Human Element, seasoned HR leader Jarlath Doherty (Squarespace, LinkedIn, Yahoo, and former CPO at a data intelligence company) shares a candid, deeply practical roadmap for evolving from a traditional HRBP/COE structure into a product-led, AI-enabled people function.

Drawing on a rare blend of engineering, data, and people leadership, Jarlath explains why EX designers and AI operations specialists are becoming must-have roles, how cross-functional pods and scrum-style sprints accelerate HR innovation, and why data quality, policy hygiene, and content governance are make-or-break foundations for AI. He illustrates what “always-on” support really means with a 2 a.m. maternity-leave example, and why employees now expect consumer-grade, personalized, immediate experiences at work.

Jarlath breaks down exactly where to start: tier-1 support and conversational analytics. He shares how to pilot quickly using MVPs, create an AI governance board across HR, Legal, IT, and Trust & Safety, and why HR’s next evolution requires a product taxonomy that maps services to AI use cases and roadmaps.And throughout, he’s clear on what AI cannot touch: coaching, judgment, empathy, conflict resolution, and human connection.

Key Takeaways

• Productize HR: Define services as products; introduce EX designers; add AI ops specialists to configure, tune, and govern AI systems.
• Establish AI governance: Stand up a cross-functional AI governance board spanning HR, Legal, IT, and Trust & Safety.
• Clean your data first: Build centralized data lakes/warehouses, consolidate systems, deprecate duplicate policies, and enforce knowledge-base hygiene.
• Start with tier-1 support: Launch a chatbot/ticketing pilot that integrates with your knowledge base; track deflection, accuracy, and CSAT.
• Adopt agile pods: Use cross-functional pods—HRBP, Talent, Ops, EX, People Analytics, HR Tech—and ship rapid MVPs.
• Build data fluency: Upskill HR teams in data literacy, auditing, and product taxonomy so they can define and sequence AI use cases responsibly.

Key Timestamps

[00:45] – Guest intro, tech-to-HR career, and why AI is truly transformational
[07:39] – The new HR operating model: product mindset, EX designers, AI ops specialists
[10:01] – Always-on experience: agentic support and the 2 a.m. maternity-leave example
[15:31] – How to organize: cross-functional pods, governance board, agile/MVP delivery
[18:05] – Foundations: tech stack reality, data quality, knowledge and policy hygiene
[25:54] – Are you data-ready? Data lakes, data lineage, duplicates, and “garbage in, garbage out”
[31:33] – Where to start: tier-1 chatbot/ticketing and conversational people analytics
[33:48] – Skills: data fluency, product taxonomy, agile mindset, and starting now

Barb (00:00)
Welcome to The Human Element, where we explore how AI and human insight are reshaping the future of HR. Today I'm joined by Jarlath Doherty, a seasoned HR executive known for leading global talent and people operations at some of the world's most recognized technology companies, including Squarespace, Calibra, LinkedIn, and Yahoo. With a background that spans computing, hospitality, and human capital management, Jarlath brings a rare blend of analytical and human-centered thinking to how organizations operate.

Today we will explore how AI is reshaping HR from operations to ethics, and how leaders can use it to elevate, not replace, the human element at work. Thank you for joining me today, Jarlath. It is so good to see you.

Jarlath (00:48)
Thank you, and thank you for having me as well. Looking forward to it.

Barb (00:50)
I know it's been a bit. Actually, Jarlath and I worked together at Yahoo a decade or more ago, which I cannot believe. It does not feel like that long, but I'm excited to chat today.

Jarlath (01:05)
Yeah, glad to be here. Glad to get going.

Barb (01:08)
You've seen HR evolve across multiple eras of technology, from digitization to data, and now onto AI. When you look at this current wave of AI, what feels truly transformational to you?

Jarlath (01:24)
Yeah, it is transformational. I think AI isn't just about changing how people work anymore. It is really changing how businesses are run from the ground up.

I think true transformation will always require re-architecting of systems, roles, and a lot of bottoms-up work to incorporate the innovation that's coming at us. When you think about how the business is reacting, that means we are going to deal with significant transformation on the HR side of the house as well.

Most significantly, as you mentioned, since the digitization of employee records, what we're learning, and the speed that we're running at, is that AI isn't just about automating tasks. It is fundamentally redefining how work gets done in HR. It is an exciting time, but there is a lot to take under consideration when you think about the HR function and how we're going to respond as well.

Barb (02:21)
Yeah, it's funny to think digitization doesn't feel that long ago. There's probably still somebody in a closet full of paper records somewhere. We should go get them and bring them into the AI era.

Jarlath (02:40)
I think we all have places where we are still using a lot of paper records. That is going to be part of the challenge when we think about the foundations we need to make AI work for us. We are all going to have to deal with that — poor data, siloed data, documents that are out of date. Those things will have to be addressed as we think about loading models and data to capitalize on the world we are walking into.

Barb (03:15)
When you think about the HR operating model and all those foundations, what do you think needs to shift first? Roles? Structure? Skills? Data?

Jarlath (03:30)
I think it is going to be a holistic shift. We have to rethink the HR operating model entirely. There are roles emerging that did not exist before — things like employee experience designers and AI operations specialists.

Employee experience is not new in HR, but the idea of a true EX designer who treats employees as users, designs journeys, workflows, and moments that matter, and applies UX and product management principles, that is new. That is where we are evolving quickly.

If we can adopt product-thinking to drive requirements, then we can go look at tooling and stack those use cases to deliver value.

Barb (04:12)
Totally. It’s such a shift. And it is funny, HR has always had this sort of service mindset, but product mindset is different — it is structured, prioritized, user-centered.

Jarlath (04:25)
Exactly. It is taking principles we know from product and applying them rigorously. You define the problem, define the user, define the workflow, and think about the moments that matter. Then you measure what is working. That is not traditionally how HR has operated.

Barb (05:02)
So if we zoom out: leaders are thinking about AI, but they are also thinking about their teams, their orgs, their culture. What advice would you give on where to begin?

Jarlath (05:20)
I think start by recognizing that AI changes expectations. Employees want always-on support. If they can get answers at midnight from a consumer app, then they expect that at work too.

We are hearing stories of employees saying, “I had a question at two in the morning, and I got everything I needed.” And someone asked, “Why were you awake at two?” And the answer was, “I sleep when the baby sleeps. I work when I can.” These are real scenarios.

Employees expect immediacy. That is the world we have to deliver into.

Barb (06:10)
Yes. Even if it is not 2 a.m., HR teams get bombarded at the end of the business day when leaders leave meetings. They wrap up their day, and suddenly the HR questions come in. Always-on support shifts everything.

Jarlath (06:40)
Yes, and many of us already have CRM-type systems with knowledge bases. That gives us a head start, because we can now load that content into models to provide the support we are talking about.

But that also means roles are changing. As I mentioned, employee experience designer, but also AI operations specialist, someone who tunes, configures, manages, and governs those systems.

Barb (07:12)
Huge shift. And it feels like the next wave of modernization for HR.

Jarlath (07:25)
It is. And again, the foundation matters. Our tech stack reality matters. Our data quality matters. That is going to make or break us. You have to be very realistic about where your data sits, how clean it is, and who owns it.

Barb (09:01)
Let’s talk about how you organize the work. Because what you are describing is not just new roles, it is a new way of delivering work. What does that look like in practice?

Jarlath (09:20)
I think we have to get comfortable with cross-functional pods. That means you have HR business partners, total rewards, HR tech, people analytics, employee experience, operations, and so on, all working together toward a shared outcome.

I think HR functions have historically felt very sequential. One team hands off to another, then to another. But this world requires parallel work. It requires agile work.

Pods let you move faster. They let you deliver value while you are still learning. They let you ship MVPs. They let you experiment, test, break things, and fix things.

Barb (09:58)
Totally. And organizations that adopt that way of working see the compounding effect quickly.

Jarlath (10:11)
Yes, and those pods should be governed. I think we are going to need AI governance boards.

So for example, HR, legal, IT, trust and safety, data security, and maybe employee communications. That group defines the rules of the road for where AI can be used, where it cannot be used, what data can go into systems, what content needs to be cleaned or updated, and so on.

Barb (10:45)
Yes. And without governance, you can end up with people experimenting in ways that do not scale or are not secure.

Jarlath (11:02)
Right. And I think the other factor is that we have to be honest about the foundations that need work. Tech stack realities are real. I have been in companies that have multiple HRIS systems globally, some acquisitions that were never fully integrated, policies that are outdated, and knowledge bases that are not maintained.

All of that must be addressed if we want AI to deliver accuracy and reliability. If outdated content gets loaded into models, the output is not going to be right.

Barb (11:45)
Are there specific foundations you have seen make or break success?

Jarlath (11:57)
Yes, a few.
One is the tech stack. What systems do you actually have? How integrated are they? Where does the truth live?
Another is data quality. If the data is messy, inconsistent, outdated, or split across geographies, then you are going to struggle.
Another is content hygiene. Do you have a central source of truth for policies and knowledge? Are there duplicates? Are they up to date? Are they written in plain language? Are they tagged?

And then there is the operating model itself. Does HR have the roles, the capacity, and the mindset to deliver in a product-and-pod structure?

Barb (12:45)
I love that you mentioned content hygiene. Nobody wants to clean their knowledge base, but you cannot do AI without doing that work.

Jarlath (13:01)
Yes. It is like moving houses without decluttering. You can do it, but you will hate yourself.

Content hygiene is critical. You cannot have 14 versions of a parental leave policy. You cannot have six different onboarding checklists in different folders.

Because when you load that content into a model, it is all going to come back out to the employee. If there are duplicates or contradictions, you will have huge problems.

Barb (13:45)
Let’s shift to data. I hear a lot of HR leaders say they are not data-ready for AI. What does being data-ready look like to you?

Jarlath (14:05)
Data-ready starts with understanding what data you have, where it lives, and who owns it. You need a data map. You need lineage. You need governance.

And then you need a central repository. That could be a data lake or data warehouse. Something that consolidates data from all the systems.

Then you remove duplicates, deprecate old data, and create consistent definitions. That is how you get to a world where models can reason over the data accurately.

Barb (14:41)
Yes, garbage in, garbage out.

Jarlath (14:48)
Exactly. And we have all seen garbage. I have inherited systems where half the job titles were nonsense, half the data fields were blank, and policies lived in someone’s desktop folder. That is real life.

But it is fixable. You need the right people, the right pod, and the right focus.

Barb (15:31)
So for a leader who is thinking, where do I even start, what would you recommend as the very first step into AI?

Jarlath (15:49)
Start with tier one support. That is the easiest, fastest, most impactful. Employees ask the same questions constantly. Policies, benefits, payroll timing, systems access, promotions, leveling, PTO.

You can load a cleaned knowledge base into a chatbot and immediately deflect 20, 30, 40 percent of inquiries. In some companies, 70 or 80 percent.

That frees up HR time for higher-impact work. It improves employee experience. And it builds the confidence and evidence you need to take the next step.

Barb (16:32)
And what is step two?

Jarlath (16:44)
Conversational analytics. Employees should be able to ask, “How many open roles do we have?” or “What is our attrition this quarter?” or “Where are we losing candidates?”

If you can connect your data to a conversational layer, you democratize insights across the business.

That is powerful.

Barb (29:45)
Because once you layer AI on top of that, the speed at which those answers are being given — whether you're feeding it the right information or incorrect information — is astronomically different than what was happening when people were having to self-serve.

So let me move us into a little speed round that we wrap up with.
What is one AI use case that every HR leader should pilot this year?

Jarlath (30:22)
It is absolutely tier-one frontline employee support, what some might call the chatbot or the ticketing service. That is where everybody should be experimenting.

For different organizations, that can stretch into the HR generalist role a bit, or into the HRBP role. It depends on how your model is set up, and where you have drawn the line around employees engaging with operations versus generalists versus people partners.

But that area is absolutely the place to begin. It is where we will get the biggest initial uplift.

I will give you a second one: people analytics. Getting that data into a data model and beginning to triage it is going to be very important. For many of us, that is actually more ready than we think, especially if you have a data lake or data warehouse already in place. That is one I would go after quickly.

There are also many recruiting use cases taking off right now. Recruiting AI specialists are moving fast. You will need to get out, play with these tools, and understand where they will impact your business. They are evolving so rapidly. The more customers they bring on, the better the capability gets. So experimentation is going to be important.

And all of this depends on where your business is and what your business is asking of the people team.

Barb (32:19)
One hundred percent. Pick a place and get started. You cannot wait it out until your tech stack five years from now is clear. That is not a thing anymore. You have to hop to it.

So what do you think is the most underrated skill for HR leaders in an AI era?

Jarlath (32:36)
The most underrated skill is data fluency.

In the world we are walking into, every member of your people organization needs to be fluent in data. Historically, you had a data scientist or people analytics partner, but I think the rigor required around auditing everything that sits in software is a skill we are all going to have to build quickly.

And also, when you adopt product and engineering principles — product-led design, user stories, agile, scrum — how do you take your current service delivery model and productize it?

If you look at your scope of services, you have to define a product taxonomy. Then you can target the workflows and begin building products around them. That is not a skill we have needed before, but the faster you can define your services as products, the faster you will get to the right AI use cases.

Barb (35:03)
One hundred percent. Last question: what is one AI tool that you personally cannot live without?

Jarlath (35:10)
I am really enjoying Claude. I am a big user. I have three running. I am always checking one against the other to see what I get from each, but I am leaning heavily into Claude.

Followed closely by ChatGPT. But Claude has been my go-to for the last few months.

Barb (36:07)
So let me try to bring us home here. It was nearly impossible to get this down to only three takeaways, but here is what I am walking away with: First: the product mindset, creating cross-functional pods, designing solutions that are user-centered, and getting your team into a motion that will work in this new era.

Second: being data-ready, remembering that information in the form of words is literally data in an AI era. Knowledge management, document structure, quality, and hygiene matter more than ever.

Third: the on-demand experience,AI being available when people need it, whether that is at 2 a.m. during maternity leave or 6 p.m. when managers finish their day. That unlock is real.

This has been such a good conversation. Thank you so much for coming on the show, Jarlath.

Jarlath (36:50)
Thank you. I really enjoyed it. And thank you again.

Barb (36:55)
And thank you to everyone for listening to The Human Element, presented by Wisq.
Follow and subscribe wherever you get your podcasts.

I am Barb Bidan. See you next time.