Once a case has been assigned for contact, six parameters decide how that contact is executed: the number we call from, the number we call, the time, the channel and its content, the intensity of attempts, and the persona that speaks. We call the way we set them a reachability strategy, and we run it as part of the service. The decision about which accounts are contacted, when and why is a different strategy, and it should stay with your analysts.

That boundary is the most important line in this article, so we will start with it, then walk through the six parameters one by one.

Two strategies, one boundary#

In early collections, a bank's analysts decide which contracts go to which channel and when. They divide the book into portfolios (cards, cash loans, instalment loans, car finance, overdrafts), into buckets of days past due, and into categories of debtor, and they assign cases accordingly. The same contract may come back in a later assignment. That is the allocation strategy, and it reflects things we do not see: the bank's provisioning, its other channels, its legal and reputational constraints and its view of each portfolio.

Our job begins once a case is assigned. Every assigned case should be touched, handled properly and described in a status the bank can act on. How we get through to the person, and in what form, is the reachability strategy. Keeping the two apart is not modesty. It is what allows each side to be accountable for what it controls.

The six parameters#

1. The outbound number#

The number shown on the customer's screen carries a reputation. It can look local or foreign, familiar or unknown, and handsets and networks increasingly label numbers as unwanted. We manage dedicated numbers per campaign or portfolio, watch how they are answered and deal with presentation and labelling problems as part of daily operations. Rotating numbers is one tool among several here; it is not a strategy on its own.

2. The destination number#

Customers often have more than one number on file. We keep, for each case, a memory of how each number has behaved, so the choice of which to call rests on evidence rather than habit. Without case-level memory, an operation can keep dialling a number that has never been answered, for years.

3. Time#

Each market has its own permitted calling hours, which we configure per client and market and the platform enforces. Inside those windows, the goal is to move away from calling everyone at the same hour as if the base were a conveyor belt. A customer who says they can talk after five has given us a signal worth more than any average, and the schedule should respect it.

4. Channel and content#

The call is the main channel, but not the only one. An SMS before the call can tell the customer who will call and why; an SMS after it can confirm what was agreed. Where calls fail, a text can be the fallback. The content of each message, and the stage of the process at which it is sent, are part of the parameter. A text complements the conversation; it does not replace it.

5. Intensity and intervals#

How many attempts, and how far apart. Intervals take account of what happened on the previous attempt, and they are tuned per portfolio. Maximum attempts are set per market and enforced by the platform. More attempts raise reach up to a point, after which they start to cost continuity: the customer who finally answers is already irritated.

6. Persona, voice and style#

Who speaks, how and in what register. Tone, pace and wording affect whether people stay on the line after the first sentence. Voice and conversation style are set per process and customer group, and tested rather than assumed. We always introduce ourselves as AI.

The six parameters at a glance
ParameterWhat we decideWhat it mainly affects
Outbound numberWhich numbers present, per campaign or portfolioWhether the call is answered
Destination numberWhich of the customer's numbers, in what orderRight-party contact
TimeWhen, within enforced market windowsReach, and the mood of the answer
Channel and contentCall, text messages and their contentReach and follow-through on commitments
Intensity and intervalsNumber and spacing of attemptsReach versus over-contact
Persona and styleVoice, tone, wording, registerContinuity and key information

Allocation strategy (which accounts, when, to whom) is not on this list. It stays with the client.

The same six, outside collections#

The parameters are the same in every outbound process; their weight changes. In lead qualification, time dominates. A person who has just asked for an offer is often talking to whoever calls first, so the speed of the first contact matters more than anything later. Persona matters next: a voice that sounds like a robot about to transfer you to nobody knows whom ends the call before it starts. Live transfers to the client's sales team are paced to its capacity and working hours, which is itself a constraint on time.

In appointment confirmation, channel and time work together against a deadline: the date of the visit. The business goal is to learn early that a visit will not happen, so the slot can be offered to someone else; a confirmation that arrives too late to act on is worth little.

In satisfaction surveys, persona carries the most weight. Reach is rarely the problem; keeping the same tone and patience on the thirty-first call as on the first is what produces more fully completed surveys. In each case, the parameters are set in the specification, tested in the pilot and adjusted in production, not copied from the previous project.

Not a stream#

The traditional way to run outbound contact is as a stream: one schedule, one number, one script for the whole base. Our working principle is the opposite. The value of each parameter can differ for each case, because people differ, and every case is a chance to use what we already know about it.

We are candid about where this stands. Several elements already run in production, across numbers, timing, channels and separate dialogue paths for different customer groups. Acting on individual signals from customers, such as a preferred time, is used in some campaigns and we are extending it. Where there is not yet enough evidence to personalise a parameter, a well-tested base strategy for the segment is the right answer.

How the parameters are tuned#

Every campaign day is a data source. Script variants, retry intervals, segmentation logic, calling schedules and voice parameters are tested in production through live A/B tests on real volume, with real outcomes deciding. Because there are no agents to retrain, a change that wins a test can be applied to every live conversation at once.

We don't manage people, so we control the service far better: one change and every live conversation already reflects it. – Dawid Wójcicki, Founder & CEO of Vee

Tuning has a rule: we optimise the reachability parameters on the rungs they influence, reach, continuity and the key information, and we report the business effect from the client's own systems. In regulated set-ups, every change to content is versioned and approved before it goes live.

The one you should keep#

There is a temptation, on both sides, to let a capable contact provider take over the allocation strategy as well. The provider sees which cases respond, so why not let it decide which cases to call? We advise against it, and we have said so to clients who asked.

Allocation is where a lender decides how to treat its customers across all channels, how to manage risk and reserves, and how to stay within its own policies. Those decisions need the whole portfolio in view and they need to be owned by the institution that answers for them. What a good provider should do is feed that strategy: return case-level statuses every day, report which numbers and hours actually produce right-party contacts, and flag records that need data review. Your analysts get better inputs; the decisions remain theirs.

Our job is to understand the client's reality, not to change it. What we can change is how effectively the customer contact is run. – Dawid Wójcicki, Founder & CEO of Vee

What to ask a supplier#

  1. Which of the six parameters do you actually control, and which are fixed by your platform or by us?
  2. Do you remember, per case, which number and which hour produce a verified contact?
  3. Are calling windows and attempt limits enforced by the system, per market?
  4. How do you test changes to a parameter, and how quickly can a winning change reach every live conversation?
  5. What will you return to our analysts so that our allocation strategy improves, without you taking it over?

A supplier who can answer these clearly is running a reachability strategy. One who cannot is running a dialler. We would rather be measured on the first, and we are happy to show how the six parameters are set in any process we run.

Follow The First Second

Occasional issues on the mechanics and economics of AI-powered customer contact. No noise, no vendor news.

Follow via RSS