AI and HR Managers: What's Actually Happening and What to Do

The honest assessment

HR has always been a weird blend of process and people. Filing paperwork, chasing references, writing policies that nobody reads, running payroll queries... and then sitting across from someone who's crying because their manager is a bully. AI is about to eat one of those halves. I'll let you guess which.

The process side is getting automated fast. Tools like Microsoft Copilot can draft employment contracts, generate policy documents, and summarise lengthy employment tribunal judgements in seconds. ChatGPT can write job descriptions that are genuinely better than most human-written ones (sorry, but it's true). AI-powered HR platforms like HiBob and Personio are already handling absence tracking, onboarding workflows, and employee data management with minimal human input. The "HR as administrator" function is shrinking.

What AI can't do is the actual human bit. Reading the tension in a room during a disciplinary hearing. Knowing that when Dave from accounts says "I'm fine" he absolutely is not fine. Making the call on whether to escalate a grievance or try to resolve it informally based on fifteen years of watching these things play out. The emotional intelligence, the political awareness, the judgement... that's not going anywhere.

But here's where it gets uncomfortable. A lot of HR manager roles are 70% process and 30% people. If AI handles most of that 70%, companies will start asking whether they need as many HR managers. Some already are. i've seen job specs where one HR business partner is now expected to cover what used to be a team of three, because "the tools handle the admin." That's the real shift. Not replacement. Consolidation.

Your exposure level: Medium

Medium means your job isn't disappearing tomorrow, but it's definitely changing shape. The administrative core of HR management is highly automatable. Policy drafting, data analysis, reporting, compliance checking, employee communications... all of this can be done faster with AI tools. If your value proposition is "I keep things running smoothly," you need a new value proposition.

The parts that protect you are the parts that involve genuine human complexity. Employee relations, organisational development, culture building, change management, coaching leaders through difficult conversations. These require reading people, understanding context, and making judgement calls that AI simply can't replicate yet. The HR managers who are thriving right now are the ones who've shifted their weight toward these areas.

There's also a real opportunity here that most HR professionals are missing. Someone needs to lead AI adoption policy within organisations. Who writes the acceptable use policy for generative AI? Who handles the data protection implications? Who manages the employee anxiety? Who retrains the workforce? That should be HR. If you don't claim that territory, the IT department will. And they'll do it badly, because they don't understand people.

The 90-day action plan

1. This week: automate your most tedious HR task. Take something you do every month that makes you want to throw your laptop out of the window. Reference request responses. Policy update communications. Absence report summaries. Paste your last three into ChatGPT and ask it to create a template. See how much time it saves.

2. Week two: test AI on employee communications. Draft a tricky all-staff email using Claude. The kind where you have to announce something unpopular without causing a riot. Compare it to what you'd write. You'll find AI gets the structure right but misses the tone. Learning to fix the tone quickly is a skill worth developing.

3. By day 30: build an AI-powered policy library. Take your top 10 most-requested policies. Feed the current versions into ChatGPT and ask it to update them, check for legal compliance issues, and simplify the language. Have a lawyer review the output. You've just done a job that normally takes months.

4. By day 45: learn AI-assisted workforce analytics. Open Copilot in Excel with your HR data (anonymised, obviously). Ask it questions like "what's the average tenure by department" or "show me absence patterns over the last 12 months." If you can pull insights from your own data without waiting for an analyst, your value goes up immediately.

5. By day 60: draft your organisation's AI use policy. If your company doesn't have one, write it. Use Claude to help draft it, which is a nice bit of meta. Cover acceptable use, data protection, disclosure requirements, and training. Present it to leadership. You've just positioned yourself as the person leading on AI governance.

6. By day 75: run an AI skills audit. Survey your workforce on current AI tool usage. Who's using what? Where are the gaps? Where are the risks? Package it into a short report with recommendations. This is pure HR strategic value and nobody else in the organisation is doing it.

7. By day 90: have the honest conversation with your own team. Which HR tasks should be automated? Which roles might change? What new skills does the team need? If you don't lead this conversation, someone else will lead it about you.

AI tools you should be using this week

• Microsoft Copilot for Work — Sits inside Outlook, Word, and Excel where you already work. Use it for drafting employment contracts, summarising long email threads, and pulling insights from HR data in spreadsheets. The Word integration for policy documents is genuinely good.

• ChatGPT for Work — Best for drafting job descriptions, employee communications, and policy documents. Give it your company tone of voice and it'll produce surprisingly usable first drafts. Also excellent for preparing talking points before difficult conversations.

• Otter AI for Meetings — Records and transcribes meetings with AI-generated summaries and action items. Invaluable for disciplinary hearings, grievance meetings, and any conversation where you need an accurate record. Cheaper than a note-taker and more reliable.

• Claude for Work — Handles longer documents better than ChatGPT. Paste in an entire employment contract or tribunal judgement and ask specific questions. i find it particularly good for comparing policy documents and identifying inconsistencies.

What to say in meetings

When the leadership team starts talking about AI and someone looks at you expecting the "people" perspective, try this: "I've been testing AI tools on our HR processes. The admin stuff is genuinely faster. But the bigger question is how we manage the workforce implications across the whole business. I'd like to lead on that." You've just volunteered for the most important project in the company.

If an employee raises concerns about AI replacing their job, avoid the temptation to offer empty reassurance. Instead: "That's a fair concern. Let me walk you through what's actually changing in your area and what we're doing about training." People don't want to be told everything's fine. They want to know you've thought about it.

When budget conversations come up: "We can probably automate about 40% of routine HR admin, which frees the team to focus on the stuff that actually affects retention and performance. Here's what that looks like." Lead with the business case. Always lead with the business case.

If the worst happens

If you're made redundant from an HR role, your skills are more transferable than the job title suggests. You understand organisations, how they work, where they break, and how to fix them. That's consulting. That's change management. That's organisational development. Companies pay good money for people who can do that from the outside.

The natural adjacent moves: management consultant (especially around people and organisational change), learning and development specialist, employee experience designer, or independent HR consultant. The freelance HR market is actually growing because small businesses need HR expertise but can't justify a full-time person. One HR consultant with good AI tools can support five or six small businesses simultaneously.

Here's something i've noticed. The HR professionals who struggle after redundancy are the ones who defined themselves by the process. "I do payroll. I do contracts. I do absence management." The ones who bounce back fast are the ones who can say "I understand how organisations work and I can help yours work better." That's a fundamentally different sell. If you've spent your 90 days building AI skills and leading on AI policy, you're in the second category. That matters more than you think.