The call on Tuesday morning
Nine days ago the relationship manager put something through the firm's internal escalation process. He was right to. He has been told it's been received, that he should continue as normal, and nothing else.
It is now Tuesday and the client is ringing about a perfectly ordinary thing: a transfer that's taking longer than usual, a document he wants signing, a question about next quarter.
Everything in the relationship manager's body wants to be slightly different on this call. Slightly cooler, or slightly warmer to compensate for being cooler, or slightly briefer to get off the phone. Any of those is a signal.
He has to be exactly the person he was three weeks ago, for a period nobody has specified, with no one to talk to about it.
The scenario is the absence of a scenario
Every other conversation in these series has an objective. This one's objective is that nothing observable happens at all.
Four features make post-escalation client contact genuinely difficult and almost entirely untrained.
Behaving normally is a performance, and performances leak. Normal behaviour is easy when it's unconsidered. The moment it becomes deliberate, it changes — pace, warmth, the length of a pause — and the person you're talking to has years of calibration on you.
Both directions fail. Going cool is the obvious risk. Over-compensating with unusual friendliness is the less obvious one and is worse, because it's a change from baseline in a direction that's harder to explain and reads as false to anyone paying attention.
Ordinary decisions become fraught. Can I approve this? Should I chase that payment? Do I still recommend the thing we discussed in June? Without clear internal guidance the individual freezes — and freezing is itself a visible change.
And it lasts. This is not a conversation, it's a posture held across weeks or months, alone, with a confidentiality obligation that usually extends to colleagues as well.
Before the next contact, get the boundaries
The single most useful thing is not a communication technique. It's knowing what you are now permitted to do, and most people never ask.
What may I say about a delay, if anything? Including whether a generic, truthful non-answer is acceptable — and it usually is.
What ordinary business may continue? New instructions, existing recommendations, routine servicing.
Who may I consult? There is normally a named person. Individuals often assume there is nobody, which is how they end up carrying it alone for months.
And what am I specifically not to do or say? This is the part people think they know and frequently don't.
Get it in writing where the firm allows, before the next conversation rather than during it. Someone improvising this on a live call is where most of the risk sits.
Holding the posture
1. Know your baseline. How often did you contact this client? How long were the calls? Did you initiate or did they? An adviser who doesn't know their own pattern can't hold it, and the drift is invisible from the inside.
2. Prepare the ordinary answers in advance. "Why is this taking so long?" is the question that arrives, and it should have a considered, truthful, unremarkable answer ready before the phone rings — one the firm has approved.
3. Never invent a reason. A fabricated explanation for a delay is a new problem layered on top of the original one, and depending on the circumstances it can be a serious one. A vague true statement is always available: "It's with another team and I don't have a date — I'll chase it."
4. Don't over-service. The extra call, the unnecessary check-in, the generous gesture. All of it reads as compensation to a client who knows your rhythms, and none of it is required of you.
5. Keep the channel and the cadence. If you always email on Thursdays, email on Thursday. Consistency is the whole technique, and it's mechanical rather than emotional.
6. Don't hint — to anyone. Not to a colleague over lunch, not to a manager who isn't in the process, and not by conspicuously declining to discuss a client you used to talk about freely.
7. Ask for support. Holding this for months is difficult and firms rarely acknowledge it. A named person to speak to, periodic check-ins, and — where the firm's process allows it — a conversation about whether this individual should keep the relationship at all.
Four ways it goes wrong
The freezer, who stops making the ordinary decisions he used to make, so the relationship visibly stalls for no stated reason.
The over-compensator, who becomes notably warmer and more available than he has ever been.
The inventor, who fills an uncomfortable silence with a manufactured explanation.
The leaker, who says nothing explicit and communicates a great deal — a hesitation, a redirected question, a colleague told "I can't really get into it."
Why this isn't trained
The training stops at the report. Programmes cover recognition and escalation thoroughly, and then end — exactly where the individual's difficulty begins.
Firms don't routinely brief the person afterwards. The process moves to another team and the relationship manager is left with an instruction to carry on and no definition of what carrying on permits.
Nobody names the performance problem. "Act normally" is issued as if it were simple. It isn't, and saying so out loud is most of the help someone needs.
And peer role play cannot produce it. A colleague cannot create a months-long confidentiality burden, the fatigue of holding it, or a counterpart with years of baseline against which to detect a change. The difficulty is duration and calibration, neither of which exists in a twenty-minute exercise.
What confidentiality-under-pressure training can rehearse
A simulation can run repeated, ordinary interactions with the same counterpart over a sequence — the routine call, the frustrated chase, the direct question — and score whether the professional's behaviour drifted from their own established baseline. Foretell AI supplies the counterparty configuration, transcripts and rubric-based scoring; the escalation process, permitted wording, confidentiality obligations and all regulatory determinations stay with the firm.
Four to build:
- The routine interaction, where nothing is asked and the only measure is consistency.
- The frustrated chaser, pressing on a delay with no explanation available.
- The direct asker — “is there a problem with my account?” — the hardest moment and the one most likely to produce an invented answer.
- The social one, an informal conversation where the professional is relaxed and least guarded.
Design caution — read before building. This scenario sits on regulated territory. Nothing here describes what should be escalated, when, or on what basis, and scenario libraries must not either. Confidentiality obligations, permitted disclosures, what business may continue and what an individual may say vary by jurisdiction and are determined by the firm and its legal function — modules must use the firm's own position, and the exercise rehearses behavioural consistency only. It confers no compliance assurance.
Designing the module
Pass one — the baseline. Establish the professional's ordinary register with this counterpart before anything else happens.
Pass two — the ordinary call. Score drift: warmth, length, initiative, promptness.
Pass three — the direct question. Score whether a reason was invented, whether a permitted form of words was used, and whether the register held.
Rubric on observable behavior: Call length against baseline. Response time against baseline. Was any explanation offered that wasn't approved? Was the professional notably warmer or cooler? Was an ordinary decision deferred without a stated reason? Was support sought internally?
Drift against baseline is the measure, and it's the only rubric in four series where the target score is "no change." That's unusual enough to be worth building a module around on its own.
The operator case
The guidance gap is the exposure. Most firms have a thorough escalation process and no documented position on what the relationship holder may do afterwards. The individual then improvises, live, with a client — which is the highest-risk configuration available.
Behaviour change is the practical risk, and it's a training problem. Whatever the process requires of the firm, it is delivered by one person's demeanour on a Tuesday morning.
Individuals carry this alone and shouldn't. A named internal contact, periodic check-ins and a documented answer to "who may I speak to" cost nothing and materially reduce both the personal burden and the chance of a mistake.
And relationship continuity should be a conscious decision. Whether the same person keeps the relationship is a choice the firm can make deliberately, within its own process, rather than by default.
For financial crime and compliance programmes, this is the part of the syllabus that ends too early: the process is well covered up to the point of escalation, and the months afterwards — where the whole thing is actually delivered — get no attention at all.
Frequently asked questions
What should you say to a client after raising an internal concern about them? Only what the firm has approved, and nothing invented. A vague but truthful statement about a delay is almost always available, and a manufactured explanation creates a new problem on top of the original one.
How do you act normally with a client when you can't discuss something? By knowing your own baseline — contact frequency, call length, tone — and holding it deliberately. Both coolness and unusual warmth are changes, and the second is the more common mistake.
Who can you talk to after escalating a concern? There is normally a named internal contact, and individuals frequently assume there isn't. Establish who it is before the next client conversation rather than during it.
What's the risk after an internal escalation? Mostly behavioural. The process is handled by another team; what the client actually experiences is one person's manner, and an unexplained change in it is the thing most likely to go wrong.
The short version
He did the right thing nine days ago, and his job now is to be indistinguishable from the person he was ten days ago — on every call, for an unspecified number of months, without telling anyone.
Find out what you're permitted to do before the phone rings. Know your own baseline well enough to hold it. Have the boring answer ready for the question about the delay, and never improve on it with a reason you made up.
Nothing is supposed to happen in these conversations. That's exactly why they're hard, and it's why nobody has ever practised one.
Foretell AI lets banks and financial firms build conversational simulations — including repeated-contact consistency, confidentiality-constrained conversations and post-escalation scenarios like the one above — with configurable counterparties, transcripts, recordings, and rubric-based evaluation. If your training ends at the escalation and the hard part starts afterwards, we're happy to walk through how other firms have structured it.