What Stuck With Me From Three Days in Louisville

A few weeks out from Marketplace Live, and I'm still thinking about the room.

We spent three days in Louisville with a group of talent acquisition leaders who came ready to work - not to be sold to, not to nod along, but to put real problems on the table and pressure-test them with people carrying the same weight. The agenda pointed at candidate fraud, candidate engagement, and talent attraction. But underneath all of it was one signal our team couldn't miss: the work is getting more complex, the stakes are higher, and the practitioners solving it together are moving faster than anyone solving it alone.

I want to tell you what rose to the top. First, one honest word about how it landed - because that word wasn't mine.

When we asked people how the event went, the number that stuck with me wasn't the nine-in-ten who said it exceeded their expectations. It was a word that kept showing up in the feedback, unprompted, over and over: community. One member described CXR with a string I won't soon forget writing, "enriching, caring, thought-provoking, energetic, safe, and no jerks allowed." Another, at her first event, said she walked out with actionable insights she could hand to her team on Monday. That's the bar. Not that people had a nice time in Louisville but that they left with something worth bringing back.

Here's what they brought there in the first place.

1. Candidate fraud isn't emerging anymore. It's here.

A year ago, a handful of companies were flagging this. In Louisville, it was the most urgent topic in the room - full stop. Fraudulent identities, AI-assisted interview impersonation, organized duplicate applications, ghost hires. One leader told the group that building her fraud response has eaten roughly half the time of two people on a five-person operations team. That's a headcount problem hiding inside a process problem.

The group worked response playbooks stage by stage, from resume screen through onboarding. What I keep coming back to is a note one member made almost in passing: authentication has to run both directions. If we're going to ask candidates to prove they're real, we owe them transparency about why, and we owe them a process that still feels human on the other side. The resource drain is as real as the risk of treating good candidates like suspects.

2. The human-AI boundary is getting drawn in the details.

We're well past "should we use AI." The leaders in Louisville were doing something harder and more useful - they were working out exactly where AI acts and where a human has to stay, function by function, role level by role level, moment by moment, in the loop. In one exercise, we split the process by voice (sourcing, operations, marketing, leadership) and asked each to draw the line for itself.

The frameworks that came out were specific, and they mostly agreed. AI owns the transactional layer: the matching and ranking, the first-pass outreach, the market intelligence, turning TA data into a story the business will actually read. Humans own the relationship: the judgment, the psychological safety, the coaching, the hiring decision, the offer. The interesting friction wasn't whether to draw the line but rather that it was the audit question. If AI is building the slate, who checks it for bias, at what cost, and how do you prove the choices were defensible? Nobody in the room pretended that's solved at their organizations.

3. Your stack is only as good as your process.

More than one leader came in sitting on real technology investment with limited adoption and messy data underneath it. One walked us through a stack with an identity crisis - good tools, under 80% adoption, a "toggle tax" on recruiters, and source-of-hire data that, in their words, "doesn't make sense."

The peer advice was blunt and consistent: you can't automate what you haven't documented, simplified, and standardized. Get the process right first. Don't reach for a small tool to solve a small problem you haven't fully defined, or you'll just solve the small thing and uncover three more. It's worth noting the most common unmet need people named all week wasn't a missing feature. It was integration. That tells you where the real work is.

4. Workforce planning gets respect when TA tells the whole story.

One of the sharpest conversations came from a leader living in a stop-start demand environment - flat hiring one quarter, then "hire 800 in four months" the next, with a mandate to "act like 911" and just be ready. Leadership wanted sophisticated answers without doing the planning work that produces them.

The peer read was that the answer isn't a better spreadsheet or a shinier tool. It's making the cost of unplanned demand visible. Build a playbook for unplanned hiring and attach a real cost to it. Post a planned-versus-unplanned metric so the inefficiency stops being invisible. And remember the part TA forgets to say out loud: the business owns workforce planning. TA plays a critical part, but when we quietly absorb the chaos and still hit our numbers, we teach everyone that the chaos is free. It isn't.

5. Authenticity wins in employer brand - especially when it's uncomfortable.

Companies navigating culture change, new leadership, and shifting EVPs are finding that polished messaging falls flat. One leader was working through exactly this: new CEO, layoffs, a new office in a new city with a real culture shock, and the honest challenge of saying something true about the company as it changes.

What the room kept coming back to: let your recruiters speak honestly. Put real people on video - hiring managers, ambassadors, actual employees. Lean on your internal voices. Candidates will either see themselves in your organization or decide they don't, and either outcome beats a bait-and-switch that shows up at the offer stage. Authenticity isn't a brand aesthetic. It's a decision to stop hiding the hard parts.

What we're already tuning

One more thing, because this community earns candor by giving it. Our members and guests told us where to tighten - more dedicated time for practitioner-and-vendor conversations, more room in the rotations to actually go deep instead of racing the clock. We heard it, and we're building it into what's next. The feedback that means the most to me isn't the praise. It's the fact that a room this generous will also tell me the truth.

If you were there, don't let the conversation end in Louisville

The exercises, the playbooks, the frameworks for where AI stops and a human starts - those don't have to live only in the memory of a room. Our Solutions members are bringing them into the Ask Practitioners forum and keeping them alive:

• What's actually working in your fraud response right now - and where is it still costing you more than it should?
• Where have you drawn the human-AI line in your own process, and where are you still uneasy about it?
• Has anyone successfully put a cost on unplanned demand and made it stick with leadership?

And if you weren't in the room this time: this is what you missed, and it's the kind of thing that's hard to recreate from the outside. The good news is the conversation is still going. These exchanges happened because TA leaders showed up ready to be honest about what's hard. That's what CXR is for and there's a seat for qualified TA leaders in the next one.