Skip the Megaproject:
Why the Best HR AI Transformations Start Small

Uphill, barefoot in the snow, both ways

For decades, HR transformation efforts (particularly with large technology mandates) have been organized around large, waterfall projects with small armies of resources across multiple workstreams.

Sometimes not-so-small armies. ATS and CRM implementations that take recruiting organizations 12–18 months to get live and stable. Benefits administration transitions with 10 or 15 phases to full production. Compensation analyses built on massive market survey datasets.

Regardless of the specific HR function or specific business problems, a massive amount of energy has historically been required to design, configure, deploy, and enable the technology that powers these efforts. And if the original objective isn't realized, that massive energy is dispersed and impact remains a dream.

There are hidden costs, too. Team fatigue, change resistance, and missed productivity during lengthy implementation cycles. The burden is real, and results have never been guaranteed.

---

Taking the builder's path

The modern AI revolution is changing what's possible.

For the first time in my decades in the HR transformation and technology space, practitioners are becoming more directly engaged in system design, highly connected technical architecture, and in many cases even shipping code. And it opens some interesting doors.

I'm a talent acquisition nerd, so I'll use examples of how the narrative is shifting in my own ecosystem. We have a recruiting technology collection that wouldn't surprise most of our peers in companies with similar size, scale, and geographic footprint. ATS, CRM, assessments platforms, sourcing tools, market intelligence solutions; the typical puzzle pieces you might expect. Nearly all of this tech is coming from third party providers, and most of those implementations were structured in fairly traditional ways.

There's nothing inherently wrong about that approach. But as my team is thinking about our goals for our new fiscal year, we're adapting how we consider what is possible.

We're fortunate to have access to a lot of the major commercial LLMs, as well as some unique and proprietary tools that, naturally, we've been experimenting with for a while now. As we've been considering which parts of our ecosystem work great for us and which bring unwelcome friction or user experiences that for various reasons don't quite nail what we need, for the first time in my career we've gone down the path to just build what we want.

And the exciting thing is that it's working.

Two ways to spend a year

The megaprojectOne track, four phases, one payoff

DesignConfigureDeployEnable

12–18 months to live and stable. Impact arrives at the end, if the original objective still holds.

The builder's pathDozens of small builds, shipped continuously

Weeks, not quarters. Each build reduces friction on its own, and the gains compound into a transformed way of working.

---

Incremental gains, compounded results

We're not building monoliths. We don't need to.

That's not to say that our group of very smart, very imaginative HR professionals has become overnight systems or software engineers. We're not going to be building any monoliths anytime soon (or likely ever again, but that's a different post). But what we're finding is that we don't actually need to be.

We aren't interested in building an ATS, a CRM, or any other kind of massively comprehensive core platform. We have fantastic partners that do a great job at that.

Those are the kinds of things that keep the big players from moving like they did when they were scrappy startups. MCPs create opportunities for us to build our own input experiences through the chat and productivity platforms that our tens of thousands of users are already deeply familiar with in their existing workflows, dramatically simplifying user enablement and adoption.

Our teams are able to see and feel forward progress and innovation at a pace that excites and even delights them, and our paradigm is shifting around how technology impacts our operational choices. Our partners in legal, data protection, and security appreciate that what we build is 100% explainable, and we're able to make decisions that align perfectly with our risk appetite, compliance strategies, and other operational obligations.

We're on pace to deliver dozens of agentic solutions this year that reduce friction, simplify things like job description creation and curation, fortify how we structure human behavior in interview structure and content management, and create consistent and fair experiences for candidates and interviewers. So many of the moments that matter in the recruiting lifecycle.

And the piece that might energize me most of all: we can quantify the impact and value for all of it.

---

There is still room for a mix

Sometimes the moon shot is the right play.

This is not to say that sometimes moon shots aren't absolutely the right call. Sometimes a big swing pays off, and we're certainly taking a few of those this year.

But organizations that find the right balance can see faster results on their AI evolutionary journey, and can keep HR teams engaged and excited about a future at the forefront of innovation in their spaces. Ultimately that's what keeps our teams energized and delivering results for the business, and I'm very excited about the opportunity this gives us for the future.