HubSpot wants AI agents running your whole go-to-market workflow, and Alec Cherrin came back from UNBOUND with a short list of features he'd turn on tomorrow and plenty he wouldn't. In a live AMA, Steve Smith and Rep Cap's resident HubSpot specialist take audience questions on Agent Hub, AI credit costs, and the new loop framework. They also get into whether HubSpot is becoming a work tech company.

Kevin Andrews has spent three decades watching HR tech vendors promise to serve small business and then drift upmarket. Now Board Advisor and Chief AI and Transformation Officer at PrismHR and Vensure Employer Solutions, he argues PEO software has held its ground because PEOs sell service plus expertise, with the technology there to back it up.

AI has made it cheap for customers to build their own tools around HR systems, which Kevin welcomes at the edges and worries about anywhere near payroll or benefits. He also looks at where small business HR tech goes next, with benefits guidance that runs year-round instead of "this 30-minute kind of panic every November."

Neil Costa started his career in e-commerce in 1996, well before "candidate experience" was a phrase anyone in HR used. Decades later, running HireClix, he's convinced almost none of them treat their career site like the storefront it's always been.

Neil argues recruitment marketing dollars are being wasted at the exact moment they should pay off. The ad gets a candidate to click, and the career site underneath it decides whether that click ever turns into an application. He points to his own informal test, asking a room of about a hundred TA leaders to raise their hand if they were proud of their career site. One hand went up. Neil ties that mostly to rebuild cycles that used to run six to 12 months, long enough that trying to fix it stopped feeling worth it.

He also lays out what a career site should actually do, recognizing a returning visitor and letting someone pick up a search where they left off, all the way through to actually applying: "Why would they, in their day-to-day lives, tolerate something that's not visually appealing?"

Bjorn Reynolds started Safeguard Global in 2008, running what looked like a global employer of record business years before anyone had a name for the category. Almost two decades later, he's convinced the AI era doesn't lower the bar for what his company needs to get right, it raises it.

Bjorn argues AI can absorb the busywork of global hiring, but it can't be trusted with the calls that actually matter, because being an employer of record means Safeguard is the legal employer of its clients' workers in every country it operates, not a vendor advising from the sidelines. He points to Safeguard's own internal test, 250 employment law questions run through the LLMs everyone's using, that came back 86 percent accurate, and the plain math behind why that's not good enough: "that 14 is what loses you your job."

He also explains why Safeguard split its payroll and EOR businesses apart, selling payroll to Deel to focus on the harder problem AI can't solve alone.

Scott Maguire spent a decade at RGP before joining Korn Ferry, where he now leads the firm's interim and project consulting practice. A year in, he's convinced the job comes down to trust, not the resume that got someone hired.

Scott argues emotional intelligence, more than technical skill, decides whether interim engagements succeed or fail, because an interim executive gets zero shared history and a fraction of the time a permanent leader has to earn trust. He points to an ERP implementation on the brink of missing go-live, saved by an all-interim team, where the client's own executive later said the fallout would have been "wildly material" without them.

He also explains why private equity firms now specifically seek out interim leaders for the highest-stakes transformation work.

One of Josh Millet's customers watched their applicant volume nearly double in a year, even while posting fewer jobs, because candidates are now sending AI agents to apply on their behalf, sometimes to roles they don't remember seeing.

Josh, founder and CEO of Criteria Corp, breaks down why a resume functions like the least trustworthy source in historical research (a first person account where the subject makes themself the hero), and why AI screening built on top of a resume just scales a bad signal faster instead of fixing it. The real shift underway isn't better filtering. It's asking candidates to show their work instead of describing it.

For years, Dean Carter turned down every opportunity to become a CEO, watching the job chew up leaders he respected. He spent that time reinventing the Chief People Officer role instead, and eventually the CEO seat came to him anyway.

Now CEO of Instill after 25 years as a Chief People Officer, including a long run at Patagonia, Dean argues that AI is absorbing the technical core of every senior role, and what's left is the one thing annual reviews and 360 feedback were never built to catch: judgment.

He gets specific about where that shows up in his own leadership, including a meeting where he only called on the loudest voices in the room. He didn't see the blind spot until it was named, and it changed who he called on the next time he ran that meeting. He also draws a hard line on what separates a real AI tool from a relabeled dashboard, and why most of what gets marketed as "AI for HR" doesn't clear that bar.

Jenn Harrold's hiring model at NewDay USA runs on a bet most of the mortgage industry won't make: skip the résumé filter that rewards prior experience, and screen for coachability instead. "It is very difficult to undo years of experience or process, or even, quite frankly, mistake," says the company's SVP of Human Resources, on why NewDay would rather train someone from zero than untrain them from somewhere else.

She backs the bet with a monthly feedback cadence instead of annual reviews, and the results show up in the numbers: NewDay's intern pipeline has doubled twice in two years, with conversion rising right along with it. She also draws the line on where the model doesn't scale, and what happens when someone doesn't take to the coaching at all.

Eric Nitzberg has delivered 360-degree feedback to executives, built from 20 stakeholders, then watched them privately negotiate with their own data before deciding what to believe. His read on why: most leaders will say the right things about wanting honest feedback, but "in your secret heart of hearts, are you really interested in the possibility that you're wrong?"

The founder of Sierra Leadership makes the case that confidence never actually improves with seniority, even for CEOs running six-figure headcounts, and explains why the hardest half of self-awareness only exists in how other people experience you. He also weighs in on where AI coaching earns its place, and where it still can't replace a human in the room.

KeyAnna Schmiedl points to a Workhuman initiative where recognition messages quietly shifted from celebrating progress to describing the work as "disentangling a ball of wires," weeks before anyone raised it in a status meeting.

The Chief Human Experience Officer at Workhuman breaks down why recognition data is a more honest, real-time signal of strategy execution than an engagement survey, why a stated company value means nothing until employees are actually recognized for practicing it, and why she built the company's approach to AI adoption around asking employees what they already wanted before writing a single policy.