I have spent the past several months facilitating the RPOA AI Impact and Innovation Taskforce, and I want to share an honest assessment of where our industry actually sits with artificial intelligence. Not the marketing version. The practitioner version.
Our Taskforce now includes 13 active members, and I am proud of how diverse that group is. We have large, mid-sized, and small RPO providers at the table. We have talent acquisition thought leaders, consultants, human capital advisors, and technology providers. We are considering adding two or three more voices. That diversity is deliberate. Our charge is to represent not just the RPOA community but the RPO industry as a whole, and you cannot do that from a single vantage point.
Our charter is straightforward: provide definition, guidance, and direction on AI for the RPO community. Part of that work, candidly, is calming fears. There is real angst among some of our members, and I understand it. We are analyzing both the short-term and long-term effects of AI on business models, on the RPO value proposition, on pricing, and on margins. We are examining how delivery models will change and how we must get ahead of that transformation rather than react to it.
A Maturity Curve, Not a Switch
One of the most useful things we have done is map where providers actually are on the AI adoption curve. It is not a binary of "using AI" or "not." It is a progression.
At the earliest stage are the laggards, providers waiting to see what happens. I will be honest, I likely have one or two of those on my own committee. The largest group, roughly 40 percent of all RPO providers, sits in early experimentation. They are trying the tools but not yet leveraging them in real delivery.
From there, providers move into functional adoption, which is growing fastest among mid-sized firms. This is where AI is genuinely being used for sourcing, screening, candidate scheduling, and candidate engagement. Those four areas are where most of the real activity is concentrated today.
The next stage is integrated, AI-enabled delivery. This is where your data architecture connects to the client's ATS, where your tech stack is woven into theirs, and where governance is in place to monitor how AI is being used. Beyond that lies what we call AI-native RPO, where AI orchestrates the process flows and human expertise concentrates on advisory and high-judgment work, with continuous process optimization built in. That is a future state. I want to be clear: nobody is there today. We see it as two to three years down the road.
The Uncomfortable Truth About ROI
Here is what we have discovered, and it is the part the industry needs to hear plainly. Nobody is fully realizing AI in their delivery models, and nobody has truly realized a return on investment from it. So far it has been a cost-plus story. Most providers are spending more on AI tools and technology than ever before, without yet being able to reduce cost.
I sat in a breakout session and listened to an employer say it directly: they do not want to pay extra for you using AI in your delivery model. The reason this is so difficult is that people have not figured out how to remove headcount through AI. Headcount has largely stayed the same. On margin compression, it has been wait and see. Some providers report a little, but prices have not changed much.
I will share one exchange that stayed with me. During one of our forums, a technology provider described something like a 3,000 percent improvement in recruiter efficiency, a recruiter going from filling a few hundred jobs a year to filling 3,000. Genuinely impressive. A buyer in that session was asked what that meant to him. His answer was blunt: he expected a 50 percent price reduction in RPO. The vendor's response was that you would simply fill more jobs and make up the difference in volume. But the counterargument is hard to dismiss. Every client has a defined number of hires. If you fill all of them at far greater efficiency, you may simply be doing the same work for half the price. That is the heart of why we are concerned about price compression and margin pressure if providers do not change their processes and their headcount models.
Resource: RPOA's AI Impact & Innovation Taskforce Early Findings
What We Are Recommending
Our executive briefings have surfaced five findings worth acting on.
First, invest in data architecture before expanding your AI tools, and build credible, independently verifiable performance benchmarks rather than relying on vendor claims that have not been replicated at scale.
Second, develop AI governance as a service, not merely as an internal control. This is a value-added, premium-priced offering.
Third, offer AI advisory services, guiding clients on which tools to use and where in the process. RPO providers should be the subject matter experts here.
Fourth, redesign the recruiter role before attrition forces the issue. Recruiters will no longer be screeners and schedulers. They will be talent advisors, workforce intelligence analysts, and client coaches.
Fifth, treat the candidate experience as a differentiator. Make the human moment meaningful and the quality of hire measurable.
We are aiming to deliver our position paper around August, and at the latest by the National Conference in Chicago. Our findings are available through the "AI and the Future of RPO" landing page on the RPOA website, free to members and accessible to non-members for a small fee.
The future of AI in RPO is still being written. Our job is to make sure our industry holds the pen.
