Riina Hellström
Founder of Agile HR Community
What if I told you that the biggest risk in HR’s AI journey isn’t the technology or the budget – but the way the teams work.
Too often, AI projects in HR run out of steam before anything useful is released. Not because the tools aren’t ready. Not because the ideas aren’t strong. But because they’re set up like HR transformation projects from 2005: rigid, over-planned, and too far removed from real users.
As a full-stack HR leader who bridges strategy, operations, and technology, I’ve seen first-hand that how you develop AI in HR determines whether it delivers real value.
Most HR leaders I speak with know the AI use cases.
Some of the top ones I see at the moment are:
-
AI Recruiters: Screening candidates, summarising CVs, and even conducting the first round of interviews.
-
AI Coaches: Helping managers have difficult conversations, prepare for 1:1s, or give feedback more empathetically.
-
Tier 0 & Tier 1 HR Support: Chatbots that respond to basic policy questions, payroll issues, and leave queries.
-
Talent Intelligence Tools: Predicting attrition, succession readiness, or learning gaps.
-
Workforce Planning Assistants: Supporting scenario modelling and headcount forecasting with real-time data.
I’ve written more use cases for AI in HR here.
What’s less clear among HR Leaders is how to actually get started, how to do it well – that’s where things tend to go sideways.
What to build isn't the hard part... How do we start?
Don’t overthink it. When your company has decided to go ahead with investing in AI technology, and you have the technical means to start, the best way forward is to dive right in and start developing with a cross-functional team.
A simple approach for an initial AI development in HR might look something like this:
Pick one meaningful use case (e.g. onboarding FAQs or leave queries)
Set up a squad with HR, tech, and someone who understands the end-user
Create documentation out of existing process descriptions, regulations, rules, policies. Teach the AI through fast iterative loops.
Launch a prototype—test it with a real but limited user group
Use their feedback to iterate and improve
Define simple product KPIs (accuracy, user satisfaction, adoption)
Keep the loop going until your AI-bot has an accuracy of answering and supporting the queries that you are satisfied with. (i.e. 97% accuracy for Tier 0 and Tier 1 questions)
However, to really reap the benefits, you need to learn Agile as a way of working. Only by doing you will learn what’s important and works in your company, not by planning it on paper.
Why Agile is non-negotiable for AI development
Agile tools are important, but for AI development in HR, the agile mindset is the real game-changer.
An AI revolution in HR demands a way of working that is fast-paced, flexible, user-centric and value-driven.
Exactly what Agile has to offer.
Here’s what Agile brings to the table:
Iterative development: Small, testable increments over long planning cycles
Feedback loops: Users test real versions early, giving input before scale-up
Cross-functional squads: HR + Tech + Legal + Analytics working side-by-side
Clear product ownership: Decisions get made—fast—by people close to the user
Rapid prioritisation: Not all features matter equally. Agile teams learn what actually delivers value
Agile ways of working foster frequent and responsive prioritisation, team-based working practices, ongoing feedback, co-creation and validation with people.
What could this look like in an HR team?
Picture an Agile HR-AI team that works like this:
A product owner in HR who understands the user and the business
A tech lead who can build iteratively and isn’t waiting for perfect specs
Legal and policy advisors embedded, not gatekeeping
A prototype out in two weeks, not six months
Feedback collected, bugs fixed, usage measured
Value delivered before you even finish your PowerPoint deck
This is how real-world HR AI teams work when they succeed.
What happens when you develop AI without Agile
The true risk isn’t AI itself—it’s building AI in the wrong way.
When HR teams approach AI development without Agile ways of working, they often end up building something no one uses. The chatbot you’ve invested in might deliver inaccurate or even biased responses, undermining trust before it gains traction. Months—even years—can be spent perfecting documentation, only to find you’re still not ready to launch. And when the tool finally meets real users and fails to deliver, confidence in HR’s ability to lead AI initiatives quickly erodes.
What happens next is painfully predictable: the team loses momentum, the budget disappears, and the business starts looking elsewhere for innovation. HR is left on the sidelines—again.
HR must lead the shift... with an Agile approach
The speed of AI development is exponential. As HR leaders, we have a responsibility to guide how these tools are introduced, how bias is handled, and how humans are kept at the centre.
The speed of AI development is exponential. As HR leaders, we have a responsibility to guide how these tools are introduced, how bias is handled, and how humans are kept at the centre.
The best way to ensure value-driven and responsible AI development is to adopt agile ways of working.
AI is already in our ecosystem. The question is: will HR be shaping it—or scrambling to catch up?
We cannot afford to sit this one out.
Let’s get practical. Let’s get cross-functional. Let’s get Agile.
Because if we get this right, AI won’t just make HR more efficient—it will make us more human.
- We can ramp up your team in Agile in less than two months through impactful training and team coaching. You’ll start leading the Agile work yourself after the first sprints or iterations. Learn more here here
