EP 134: Is Your HR Actually Ready for AI? The 5-Question Readiness Test | AI in HR Series Ep. 1

Episode Summary

“Should we be using AI in HR?” is the wrong question. The answer is yes — and honestly, you already are, whether you’ve named it or not. Every template your payroll software spits out, every auto-declination email, every benchmarking prompt is a version of AI that’s been running in HR for years. The real question is sharper and more useful: how ready is your HR function for AI? Because AI multiplies whatever foundation it’s sitting on. Clean data, documented process, and a human owning the final call, and it becomes a genuine force multiplier — the kind that let Kerri run her consulting practice with an intern instead of two employees and nearly double her revenue. But scattered data, no documented process, and no rules, and AI will make expensive mistakes faster than you can catch them. Messy data in, messy data out. This episode kicks off a 12-part AI in HR series, and it starts exactly where you should — with a readiness audit. Five questions to run on your own HR function before you roll out a single tool. The tool isn’t the strategy. Your foundation is. AI just makes it bigger.

In This Episode

• Why “should we use AI in HR?” is the wrong question — and the better one to ask instead

• The 5-question readiness test to audit your HR function before you add AI

• Why a documented process matters more than whichever tool you pick

• How dirty, scattered people data turns into expensive AI mistakes at scale

• Why every business needs a simple, one-page AI use policy — whether you’re in HR or not

• The human who has to own every real HR decision, and why AI never replaces that judgment

• The stats: 54% of organizations use zero AI in HR while 92% of leaders expect more this year

• Real examples of Kerri using AI for turnover modeling, compensation studies, and taming her inbox

Chapter Timestamps

0:00 The wrong question everyone asks about AI in HR

2:20 Why we’re starting with a readiness audit

4:41 Question 1: Is your process documented or in your head?

9:25 Question 2: Do you know your core people numbers?

11:44 Question 3: Is your employee data clean and in one place?

14:09 Question 4: Do you have an AI use policy?

15:30 Question 5: Who owns the final HR decision?

18:57 Why AI readiness is a financial issue

21:20 The human behind the AI (the gas-pump rule)

24:00 The stats: what’s really happening with AI in HR

28:21 How Kerri uses AI with clients

37:00 Recap + your action item

Resources Mentioned

• The HR Easy Button (book): saltandlightadvisors.com/thehreasybutton — paperback, hardback, and Kindle on Amazon

• HR Foundations (course, under $400): saltandlightadvisors.com/hrfoundations

• Free Mini HR Audit (10-minute quiz + score): saltandlightadvisors.com/hraudit

• HR coursework, tools, and templates: saltandlightadvisors.com/resources

• Companion blog post: Is Your HR Actually Ready for AI? — saltandlightadvisors.com/blog

Your Action Item

This week, run the five readiness questions on your own HR function — out loud, with your leadership team if you have one. Be honest about where the process lives only in your head and where your data is scattered. Then take the free Mini HR Audit at saltandlightadvisors.com/hraudit. It takes about 10 minutes, gives you a score, and shows you exactly which foundation to fix first — before you make anything fancy with AI.

GET THE MONDAY EMAIL

If this episode resonated, you’ll like the Monday morning email. Every Monday at 5:28am, Kerri sends one practical idea for leaders who want to do this work better — including the conversations behind episodes like this one. 1,000+ leaders. 50%+ open rate. → saltandlight.myflodesk.com/saltandlightadvisors

Full Transcript

The wrong question everyone asks about AI in HR

I’ve heard so many people ask, should we be using AI in HR? And that’s the wrong question, in my opinion. It’s a resounding yes, you should be using AI in HR. But what I would say is, how ready is your HR function for AI? Regardless, yes, there are certainly ways that I want to see you use it.

So welcome back to another episode of Don’t Waste the Chaos. I’m your host, Kerri Roberts. I’ve been doing human resources for over 20 years. I’ve been utilizing AI in human resources — in the terms of what you’re thinking — for three years. But we’ve been using AI with HR payroll software, with the Society of Human Resource Management, with some of those websites, for a really long time, because we’re prompting a question and it’s spitting out a template. The reporting features inside HR payroll technology, where you put in that you’re declining a candidate and it automatically sends a declination with templates and language built in — this has been going on. But the way you’re thinking about AI right now, I’ve been doing for three years.

I started with ChatGPT, and now I use ChatGPT and Claude, some Gemini, a few other tools. I’m not here to promote one specific tool. I’m here to talk about AI in HR in general. And I’m kicking off what I think is going to be a 12-part series. I just did what I was calling a 13-part series on HR foundations, and it ended up being 14 episodes because I had a little bit more to share. So I’m going to say this is a 12-part series. When I started thinking through the aspects of artificial intelligence in human resources — the implications, what to consider, the questions we need to ask, how to implement, what the policies might look like — it became long.

Why we’re starting with a readiness audit

We’re starting this episode with a readiness test, a little bit of an audit, because I want to ask you some questions to get you thinking about how you’re already using it and what your HR foundation looks like. I would be floored if you’re not using it at all in any capacity right now.

I listened to a podcast episode yesterday of a woman in a high-power position at a large, innovative company. She’s 40 years old, and she’s not using AI at all. It surprised me, because I was an early adopter — a little bit of an enthusiast. And yes, there are security issues that come up and things you need to do proactively to safeguard yourself. I’m not saying willy-nilly throw everything into HR work. But in general, it’s where it’s at, friends.

When I look at my consulting revenue and how it’s helped me: I don’t have any full-time or part-time employees anymore. Right now I’ve got a summer intern, and that’s it. This time last year, I had two part- to full-time employees and a summer intern. So it’s saving me around $70,000 annually just in people alone. And if you’re listening thinking, “Thanks, Kerri, I’m an HR person and you’re talking about eliminating me” — no. I didn’t eliminate myself. I’m the HR consultant. I streamlined and automated how I come to market, and my revenue is going to almost double this year. I’m automating my admin, my marketing. There are things to automate that don’t replace you. They make you far more efficient and effective.

I love AI. I use it every single day in my business and for my clients. So when someone asks, “Should we use AI in HR?” — yeah. But before you get into that, let’s run this little assessment to see how AI-ready your HR function already is. Here’s the five-question readiness test.

Question 1: Is your process documented or in your head?

Is there a documented process that AI can actually learn from, or does it all live in your head? If it all lives in your head, that’s okay. You can click that little microphone button on your phone or computer and dictate to AI, and it can help you write standard operating procedures. That shouldn’t deter you.

Is AI going to help with your HR function right out of the gate? No — but it’s going to help you standardize your processes so it can make an impact. This is no different than when I have clients implement HR payroll software. When the implementation analyst asks, “What are your holiday policies? Your paid time off? Sick leave? Bereavement? Unpaid time off? What does FMLA look like — does your company even qualify based on size? What about ADA leave, personal leave, sabbaticals?” — if you don’t know the answers, the analyst can’t set up your system well. Same with benefits enrollment, performance reviews, applicant tracking. They can recommend, but they need the structure from you.

Here’s the difference with AI. A payroll vendor can recommend but can’t decide for you. With AI, you can say, “I don’t have a process for this. Can you benchmark best practices for organizations in the state of X, in X industry, around X in revenue, with roughly X employees, and give me recommendations that are legally compliant at the state and federal level that I can make a decision on?” Then it gives you the data and its recommendations, and you still decide: “This is right for my business and my culture,” or “It’s not, and here’s how we’d change it.” AI gets you further because it can do the research and give you a starting point. But if you don’t have the foundation built, we shelve the HR piece and first document the processes with AI.

Question 2: Do you know your core people numbers?

Do you know your core people numbers? I’m talking about turnover, cost per hire — all of it. That data is useful because if AI doesn’t know how you’re operating, it might lead you astray. It’s good to feed your people data into your AI so it’s answering based on an organization that operates like yours.

If you ask, “What’s a good employee reward and recognition program?” it’s going to give you best practice — probably something Google does. But if it doesn’t know you’re on a shoestring budget, that you did a layoff two years ago, that your turnover is 92% — it’s going to lead you astray. A lot of folks aren’t measuring HR metrics, which, by the way, are finance and business metrics. They’re not measuring them because nobody owns it. So they don’t know what data to feed AI so it gives accurate responses based on how they actually operate.

Question 3: Is your employee data clean and in one place?

Is your employee data clean and in one place, or is it scattered across inboxes and OneDrive and people’s desktops? The beautiful thing: AI can help you clean it up if you give it access — and not unfettered access. I make it ask me to grant permission every time I want a project done, instead of just handing it the keys.

But if your data isn’t together, now your project isn’t “AI and HR” — your project is data and record management. And that’s okay. If your HR documentation is spread across people — your CFO has some, you have some, your office manager has some, your benefits broker captures some you never pull down, your 401(k) provider has more — and you don’t have a good record-retention strategy, that’s the first project. You can’t do much with information you can’t find. Once it’s streamlined into a source of truth, then we can start reading it and making it actionable.

Question 4: Do you have an AI use policy?

Do you have any rules about what your team can and cannot put into AI? A simple one-page policy counts. If you don’t have rules, two things happen. Some people won’t use it at all — they’re uncomfortable, frustrated, or a slow adopter. And a lot of people think the danger is only in the lack of rules. But the other side is: if you don’t have rules and standards, it doesn’t get utilized. Just like any technology rollout — a training platform, a CRM — adoption varies. Some adopt, some overuse it, some never adopt.

So what’s the policy? And don’t forget the training piece — that’s huge, and it’s a place where HR should make an impact through learning and development. Know your expectations, your policies, how people get trained, and what accountability looks like. And FYI: you already have people using it. Get that policy going right now, whether you’re in HR or not.

Question 5: Who owns the final HR decision?

If AI drafted a decision tomorrow that’s an HR decision, is there a human who owns the final call? This is a big one. Here’s how I know it’s a problem. I have around eleven clients right now, and I’d say nine out of eleven don’t have one HR decision-maker other than me. Usually there isn’t one internal person who owns it. Even the business owner shows trepidation about who can make the call.

The same happens with AI. AI will say, “In your state and sector, here are recommendations, benchmarks, and data with sources cited.” But you still have to say, “That’s appropriate for my business, we’re going to implement it, in this timeframe, rolled out this way.” Somebody still has to say it. So — who owns HR decisions at the end of the day?

That’s your five-part HR AI readiness check. Reflect on those, talk about them inside your business, and see where you’re at. I’m focused on HR, but you can apply these roughly to any department.

Why AI readiness is a financial issue

Is readiness a financial issue? Absolutely. We’re talking about dirty data, time spent, consistency, and the potential for discrimination if we go down the wrong path because we haven’t fed AI the right data — and it makes the wrong recommendation. An AI tool trained on messy data makes expensive mistakes at scale. It might point you toward bad hires, mispriced offers, or a policy gap that becomes a settlement. HR metrics are finance metrics, and AI multiplies whatever foundation it’s sitting on, good or bad.

Say you put in your compensation data but miss the philosophy behind it — the last time you did raises, why you pay at a certain level — and AI starts recommending increases benchmarked to Silicon Valley when you’re in Ohio. There are real financial implications if you don’t do your AI integration correctly. Messy data in, messy data out.

The human behind the AI (the gas-pump rule)

There’s an obvious need for humans in your business. When the big players talk about “50 AI agents,” what they really mean is they have human employees, and then AI agents as subordinates to those humans. That’s part of the policy: how are you going to treat this?

I have projects, scheduled tasks, and a daily dashboard. I used to live off my calendar; now I live off my AI dashboard. Once I finish recording this episode, there’s a task waiting to take the transcript and put together my Monday elements: draft an email about my podcast, draft show notes for my production company, and draft two blog posts — one for my personal website and one for my Salt & Light Advisors website.

But I read all of those. I’ve trained it on my voice — I gave it the full manuscripts of my books, a ton of emails, and my transcripts, which is exactly how I talk. And there are still things it does where I go, “That’s not how I talk,” and I give it feedback. I am a huge part of all of this. We’re not outsourcing to AI entirely. I’m now doing tiny portions of my old admin’s and marketing person’s jobs, because I’m the decision-maker. It actually added a little back to my personal plate.

So think about it: what human manages the AI? You can’t just turn scheduled tasks on and let nobody watch them. I don’t have my blog posts go live — I have them saved as drafts, and I read them, review them, and choose when they publish. You’re ready for all of this when you have clean data, documented processes, a simple use policy, and a human owning every real decision. Once that’s in place, AI is genuinely a force multiplier. But it never owns judgment, coaching, or a call that affects someone’s livelihood. That’s still on you. That’s still on me.

It’s like pulling up to a gas station. The pump does a lot, but you still have to unlock the tank, unscrew the cap, choose your gas, put the pump in, and put it back. Same deal. There are things that stay human.

The stats: what’s really happening with AI in HR

I did some research on stats, and you don’t want to miss these. 54% of organizations still haven’t adopted any AI of any kind in human resources, with no immediate plans to. But 92% of executive leaders expect more AI in the workforce this year. If it’s happening in other areas of your business, you have to put it into HR — because HR is the front door.

If you’re representing why a person would want to come to your organization, but you’re not using any of the tools that person will be hired to use, that’s weird. It’s the same reason I harp on not doing manual benefits enrollment and manual onboarding. If someone’s coming into a tech role, expected to use a CRM every day, but they had to fill out paper to get started — that’s an oxymoron. It makes you look silly and it’s not aligned, which speaks to your culture right then and there.

That 54% figure is from SHRM’s State of AI in HR 2026, which surveyed a couple thousand HR pros. Another one: only around 30% of organizations have any AI guidelines or policy in place. And AI use at work — not just in HR — jumped from 21% to 40% of employees in two years. Daily use doubled. It’s already in your building, ready or not.

Another piece of data: replacing one employee costs one-half to two times their annual salary. Voluntary turnover costs US businesses around $1 trillion a year, and 52% of that is preventable. That’s where using AI on your HR metrics gets powerful. If you’re not tracking absenteeism or PTO utilization, you probably don’t know that person has burned all their PTO over six months, which often signals they’re getting ready to quit. You have access to a ton of data if you have HR payroll software or even spreadsheets. But if you’re not analyzing it, AI can take that data, run analytics, and tell you where to look. You’ve just got to have clean data and access to it.

The last stat: only 15% of workers think AI will eliminate their job within five years. When I started thinking about this episode, I didn’t want to scare anybody. What’s trending on YouTube is “AI is going to take your job.” It’s real, but it feels scare-tactic-y, and that’s not my style. Is AI going to take their jobs? No. Is it going to impact their jobs? Absolutely. So we all need to keep an eye on it.

How Kerri uses AI with clients

I want to highlight a couple of HR ways I’ve used AI with clients. First, turnover data. I fed AI a client’s headcount, average salary, and separations, and it built a model that told me what it actually costs them when they lose an employee. We say it costs half or two-thirds of a salary — but I want to know what it costs in your business, so the next time a manager says they don’t have time to properly train or coach someone, I have data: “You’re costing the company this much when you don’t invest in your people.” I decide what to include and what it means; AI helps me analyze it.

Second, compensation studies. Clients give me a spreadsheet with every employee’s name, start date, current salary, exempt or non-exempt status, date and amount of last pay increase, and why. I have software that pulls the 10th, 25th, mean, 75th, and 90th percentiles based on geography, revenue, employee count, and industry. I add those columns and plot each person. Then I upload the spreadsheet to AI and ask, “Based on this data, what trends do you see? Any red or yellow flags?” It helps me hone in and sparks the conversation with the client. Then I take what I learn back to AI and ask for recommended next steps. It shortens my turnaround.

But my favorite way — if you’re not using it for this, it’s a game-changer. Go to your inbox, to your oldest email. Copy it into AI and say, “I’ve let this sit for a while and I need to respond, but it’s taking me time. Help me think through this so I can start mapping a response.” Or, “What questions do you need to ask me to help me develop a response?” I do that all the time.

Here’s an example. Someone asked me to be featured on MSN — a great platform — but it cost $500, and I never pay to be featured. So I opened Claude and said, “Here’s the opportunity. I generally say no to paying for promotion — I’d rather use that money elsewhere. Is this where I want to start? Help me think through whether it gets me leverage.” Claude laid out the potential reach, and recommended I not do it — I could put that $500 into ads and get more. It helped me think it through when it had just been sitting in my inbox, and it pulls in research I don’t have or don’t even know how to Google in the moment.

Recap + your action item

Quick recap of the readiness check: Is there a documented process AI could follow, or does it all live in your head? Do you know your core people numbers? Is your employee data clean and in one place? Do you have any rules for how your team can or can’t use AI? And if AI drafted a decision tomorrow, who’s the person who says, “Yes, that’s accurate,” or “No, we need to change it”?

I recommend you build your HR foundation before you make it fancy with AI — and AI can help you do it. To see what your foundation is currently tracking at, do my free Mini HR Audit at saltandlightadvisors.com/hraudit. It takes about 10 minutes and gives you a score. I also have affordable coursework at saltandlightadvisors.com/resources, and my book, The HR Easy Button, is on Amazon.

This is Episode 1 of the series — eleven more are coming, digging much deeper into how to use AI in your HR department, including the policy piece. If you found this helpful, share it, subscribe, and leave a review — it helps the algorithm push the show up. Implementing AI in any area of your business, especially HR, can feel chaotic — but the work is worth it. So don’t waste the chaos. Embrace it. I’ll see you next week.

Resources To Keep Building

Before you add AI to your HR function, take the free HR Audit to see whether your people systems are actually ready for it — in about 5 minutes. → saltandlightadvisors.com/hraudit

🎯 Take the free HR Audit — Score your HR systems in 5 minutes and see exactly where your gaps are. 👉 saltandlightadvisors.com/hraudit

📚 Explore the HR coursework — Self-paced courses, tools, and resources to build your HR systems step by step. 👉 saltandlightadvisors.com/resources

✉️ Get the Monday Email — One practical idea for leaders, every Monday at 5:28am. 1,000+ leaders, 50%+ open rate. 👉 saltandlight.myflodesk.com/saltandlightadvisors

Need fractional HR support or want to talk through a specific challenge? 👉 saltandlightadvisors.com/contact