AI agents are already handling work across millions of rental units at Get Covered, and the company’s chief operating officer says the hard part is deciding who is responsible for them. Rick Folgmann’s practical advice is to write down each agent’s rules, match its freedom to the workflow, and assign a person to review its work.
Watch Desk analysis
What happened
Folgmann says Get Covered runs 16 AI agents across more than three million rental units. In a guest article for Crunchbase, he argues that the key questions are managerial, not just technical: what decisions an agent may make, how its work is checked, and who answers when it gets something wrong.
His examples put boundaries around autonomy. The company can send a routine reminder to a vendor automatically, he says, but does not automate responses to residents requesting emotional support animals. He also recommends starting with tasks governed by written rules, so teams have something concrete against which to check an agent’s output.
Why it matters
The advice gets at a practical distinction often lost in the rush to deploy agents: the same system need not have the same freedom in every task. A low-cost mistake spotted quickly may justify more autonomy than an action with serious consequences or a long delay before anyone notices.
Folgmann says each agent at his company has a named person responsible for reviewing its error log, adjusting its autonomy and answering for its actions. That offers a clear test for organisations adopting agents: if nobody owns the workflow, “the model did it” is unlikely to be a satisfying explanation.
Our read
This is useful operational advice from someone whose company says it is running agents at substantial scale, not a universal blueprint or independent evaluation of the system. The strongest point is also the least glamorous: write the rules down, set boundaries task by task, and give each agent a manager with a name. Autonomy without accountability is just a faster way to lose track of who was meant to be in charge.
What to watch
- Whether other organisations publish concrete rules for assigning responsibility to agents.
- How companies decide which workflows can run automatically and which need human approval.
- Whether teams share evidence that named owners and routine error-log reviews improve outcomes.
Discussion spark: Should every AI agent have a named human owner who is accountable for its actions, or can responsibility sensibly be shared across a team?
Sources and evidence
- Source update (27 October 2026, 11:00 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.