At Gartner HR 2026, SeekOut President John Tippett started with a simple observation. Every role in talent acquisition is really two jobs. There is the job you were hired for: judgment, persuasion, relationships, advising hiring managers. And there is the job around the job: reporting, scheduling, status-chasing, research, and data entry across systems.
Nobody loves the second job, and it is not small. A survey of 748 HR leaders found 17.7 hours of manual admin for every open role, more than two full days. Those tasks look exactly like the things AI should handle. And yet more than 65% of recruiters already use AI somewhere in their workflow. So why hasn’t it given them the time back?
Recruiters became the glue between their tools
The average TA team runs on a stack of HR and recruiting tools, before you count email, Slack, and the calendar. The recruiter ends up as the bridge between all of them: retyping intake notes into each tool, copying shortlists into the ATS, chasing hiring manager feedback across threads. Sprinkle AI into any one of those tools and the recruiter’s week does not change, because the work between the tools is still theirs.
Onboard an AI agent the way you would onboard a new hire
John’s answer is to treat an AI agent the way HR already treats a new employee. Not because agents are people, but because they need the same things to succeed.
- The right tools and the right access. You would never hire a recruiter and withhold the ATS. An agent without access to email, the ATS, Slack or Teams, and the calendar becomes another silo, and the recruiter connects everything by hand.
- Work in the right place. More and more people work inside their company’s AI assistant, whether that is ChatGPT, Claude, Gemini, or Copilot. An agent that lives in a separate destination becomes an island nobody visits.
- Room to iterate. Even the smartest new hire has to learn how the company works and take feedback. An agent has to do the same: ask questions, take direction, and change what it does in response.
- Autonomy earned over time. Nobody sends a new hire to pitch the biggest customer on day one. An agent should start with review and oversight, and take on more on its own only as it proves the work is right.
What you never delegate
Just as important is the list of things that stay with people. In hiring, John argued, it is everything in that first job.
- The hiring decision. Accountability stays with a person. Also watch the perception: a message that lands at 2 a.m. can look like an AI made the call, even when a person was driving.
- Persuasion and closing. Candidates change jobs for people they trust, not for well-formed emails.
- Advising hiring managers. Agents bring the data, but recruiters make the argument.
- Moments of trust. Rejections after final rounds, bad news, a candidate’s hard career decision.
“If you onboard an AI agent like you would onboard a person, give them the tools and access, allow them to work where the work is happening, make sure they can learn your systems and iterate, and that they earn autonomy over time, we think all of that second job can go to AI systems.”
John Tippett, President, SeekOut
What it looks like in practice
In the second half of the talk, John walks through how SeekOut applies these ideas. It runs inside the AI system the company has already approved, so it inherits the connections IT has already made to email, the ATS, and Slack, with no new integrations to request. It builds data-backed artifacts like market reports. When you ask for a sourcing plan, it asks clarifying questions instead of guessing. It checks with you before it acts. And because chat alone is like a restaurant with no menu, it pairs the chat with a visual UI that shows what needs you today.
See it on one of your roles
Want to see an agent do the job around the job on your own reqs? Book a 30-minute working session.
Recruiters direct. Agents execute. Your rules govern.
Stat sources: Totaljobs (The Stepstone Group), survey of 748 HR leaders, Aug 2025; Employ, Recruiter Nation 2025.
