AI answers the copay question. A person owns the coverage call.
A Brisbane coaching founder, a US benefits founder and an Oslo insurance agency owner agree on what AI took, and on who still answers for the rest.
Ask a benefits bot your ER copay and you get the answer in seconds, with the plan page cited.
Ask who is accountable when that answer is wrong, and you've asked a different question.
Three leaders in Australia, the US and Norway. Three industries. One line.
Nobody here is anti-AI. They handed over real work, and they can tell you exactly what they kept.
We asked which parts of a job AI has actually taken, and which still need a person. In financial coaching, benefits and insurance, the answers rhymed. It matters now because every company is deciding, role by role, what to hand over and what to keep.
The factual tier is gone. Fully.
Rick Ramos, Founder of Benefits Studio (United States)
“The factual tier has gone fully AI: "What's my ER copay?" or "Is this provider in-network?" Benny handles those in seconds, citing the exact plan page - faster and more accurate than digging through a 120-page PDF.”
“The most expensive mistake I see companies make isn't automating too much or too little - it's automating the answer while forgetting the accountability.”
A US benefits founder hands the lookups to AI without a fight, because speed on facts is a real win. Then he points at where the bill lands when an answer turns out wrong, and who is standing there.
Intake is automated. Inspections aren't.
Mikhail Kovalev, Agency Owner of Kovalev Insurance Agency Inc. (Oslo, Norway)
“Automated systems can easily handle initial policy intake and baseline replacement cost estimates.”
“However, evaluating nuanced property inspection conditions and tailoring commercial liability exposures still demand experienced human counselors.”
Insurance Agency in Norway follows the same script: software takes the first pass, and experienced people take the hard calls. His rule for choosing both is short, and it is about stakes rather than headcount.
The tool spots the overspending. It can't say why.
Rebecca Maher, Founder and CEO of My Money Circle (Brisbane, Australia)
“A tool can show someone is overspending. It can't help them understand why it keeps happening, or hold them accountable to change it.”
“Leaders should ask where trust is earned in their customer relationship, then automate everything around that moment so their people can spend more time in it.”
A Brisbane coaching founder has automated content, bookings and cashflow mapping, so the tools get the spreadsheet work. Every member's numbers still get a human check, and the coach gets the conversation.
Where they split
AI took: benefits Q&A, policy intake, baseline replacement estimates, cashflow mapping, callback bookings.
People kept: the why behind the overspending, the property inspection, commercial liability.
Nobody here wanted a human on everything. Nobody wanted a bot on everything, either. The line sits where a wrong answer lands on a real person.
My take
Before you decide human, agent or both, split the role into tasks that have an answer and tasks that have an owner. Ask of every task: if this is wrong, who answers for it? If the answer is a person, that task is part of a job, not a line in a prompt.
Your turn
In your business, where does trust or liability land, and who owns that moment today: a person, an agent, or both?
Next: an agency CEO letting agents do most of the work asks where young people will learn, while a governance consultant and an executive coach say junior reps must be redesigned.
