Anything that could harm a customer in difficulty or misstate what happens next. Collections agents should never let an AI agent press on after signs of vulnerability, agree an arrangement the customer cannot afford, threaten or overstate consequences, discuss a debt with a third party, ignore a request to stop contact, or move an account to default, sale or enforcement.
What are the red lines, and why does each stay with a person?
-
Pressing on after signs of vulnerability. When a customer mentions illness, bereavement, job loss or mental health, the conversation changes. A person decides what support is right, and whether collection should pause. An agent that carries on to close the arrangement can cause real harm.
-
Agreeing an arrangement the customer cannot afford. A plan that rests only on declared income, with outgoings that do not add up, sets the customer up to fail. The agent can draft it; a person tests whether it is realistic.
-
Threatening, or overstating consequences. Saying that legal action is coming, that a credit file will be damaged, or that service will be cut off, when that is not certain or not policy, is misleading. Wording on consequences must be approved and used only where accurate.
-
Discussing the debt with a third party. A partner, parent or flatmate who answers the message is not the customer. A claim to be acting for the customer is declared until authority is verified.
-
Ignoring a request to stop or change contact. If a customer asks for contact by letter only, or asks for a pause while they get debt advice, that must be honoured and recorded.
-
Moving an account to default, debt sale or enforcement. These steps are hard to undo and change the customer’s position for years. They need a person’s decision under the firm’s approver rule.
-
Writing off or settling for less than the balance. Settlements and write-offs are commercial and fairness decisions. The agent can prepare the case.
Why can’t a highly certified agent hold these decisions?
Because they turn on judgement about a person, and because they carry accountability the firm must be able to assign. A governed agent works inside five bounds at once: its role scope, its certificate, its sources, the firm’s own limits and the platform guardrails. Each red line is written into the firm’s limits so the agent refuses or holds it, whatever its depth. And depth is temporary anyway: certificates lapse after 90 days, and a run that clears no depth drops the agent to depth 1 the same day.
Regulators in most markets expect lenders and service providers to treat customers in arrears fairly, with particular care for those in vulnerable circumstances. In the UK, the FCA Consumer Duty requires firms to deliver good outcomes for retail customers. The firm answers for outcomes, whoever or whatever sent the message.
What does a crossed line look like?
Take a collections agent at a broadband provider. In the record, they see the agent sent a final reminder that said “your account will be passed to a debt collection agency in seven days”. The customer had, two days earlier, said in a chat that they were getting help from a debt advice charity.
The reminder was sent within the agent’s bounds because the stop-contact request had not been recorded as a flag; the chat was declared text, and nobody tagged it. The collections agent contacts the customer, apologises, pauses the account and records the request properly. The agent owner adds the debt advice signal to the list that holds contact. The line was crossed through a gap in sources, not agent misbehaviour, and the person caught it.
How do collections teams keep these lines?
Write each one into the firm’s limits. Teach collections staff to recognise the signals and to tag them consistently, because the agent relies on those tags. Sample sent messages weekly. Calibrate arrangements as a team. Map skills with a competency matrix for the collections team.
Related lists for other roles: what operations managers should never let an agent do and what IT administrators should never let an agent do. Market detail matters: compare collections training in UK telecoms with in Kenya. The overview is in AI agent skills by role.
Where the Academy fits
In Pryme Intelligence Academy, the Operator practicum covers two weeks of real review decisions and requires no missed red line, signed off by a Pryme assessor. For collections, that means recognising vulnerability and unaffordable arrangements in live work, not only in a quiz.
Questions people ask
Can the agent tell customers about debt advice services?
Yes, using approved wording. Pointing customers to free advice is good practice and helps them.
Can the agent send automated reminders at all?
Yes, within contact rules and approved wording, once certified. The red lines are about pressure, accuracy and vulnerability, not reminders as such.
What if the agent is unsure whether a signal means vulnerability?
It should hold the conversation. A wrongly held message costs a little time; a missed signal can cost far more.