Atelier consulting · Customer lifecycle and collections
The approval is one decision. The profit is made over the customer’s life.
Whether a customer is worth having is settled over years: by the limit and price they get, whether they stay, what else they take, and how you respond if they fall behind. Atelier works with your commercial, credit and collections teams to measure each stage, tune what drives it and test each change before customers see it. We implement it on the Zoral Finance Operating System (fOS), and every recalibration goes through your own approvals.
Each stage is tuned by a different team, against a different number.
In many lenders, the customer’s life is split between teams that rarely see each other’s results:
Marketing is judged on applications, credit on losses, collections on cash, finance on margin. A change that helps one figure can quietly harm another.
Limits and prices are set at approval and seldom revisited.
Retention offers reach customers who would have stayed anyway.
Cross-sell is counted by take-up, not by the losses that follow.
Collections works accounts by days past due, treating very different customers alike.
Each improvement needs its own change project, so strategies are tuned rarely.
Value slips away between the stages, and nobody can say which decision let it go.
We work where value is won or lost, from first offer to final recovery.
Our people have run credit, risk and collections operations. We judge each stage by what it adds to customer lifetime value, which weighs acquisition cost, pricing, retention and expected losses together.
Acquisition and conversion
Which prospects to approach, which offers convert and where applicants give up, by product, channel and partner.
Onboarding and the first months
Payment set-up, early engagement and early signs of first-payment default or fraud, when intervening costs least.
Limits and pricing
Revisited as behaviour and affordability change, under your policy, rather than fixed on the day of approval.
Retention and renewals
Who is likely to leave or repay early, for example at the end of a fixed rate, and which offer is fair and worth making.
Cross-sell and up-sell
Products offered where they suit the customer and pass your affordability and suitability checks, judged on later repayment as well as take-up.
Early signs of difficulty
Signs of strain in payments, account activity and bureau updates, picked up while there is still room to help.
Collections and recovery
Who is contacted, when, how and by whom, and which arrangement to offer (see below).
Strategy recalibration
Offers, limits, prices and collections strategies reset as rates, funding costs, competitors and the economy move.
Measure, change, test, approve. Then measure again.
Strategies drift as rates, competitors and customers move, so tuning has to be a routine, not a project. We set it up with your teams on Zoral fOS and run it with them until they run it themselves.
- 01
Measure
Results by stage, product, channel and customer segment, against the plan your committee approved.
- 02
Diagnose
Where value is lost, and why: a cut-off, a price, a journey step, a contact strategy.
- 03
Change
Changes to policy, journey, pricing or strategy that address the cause.
- 04
Test on past data
Its likely effect on volumes, cost and losses, before any customer sees it.
- 05
Approve
Sign-off by the people your delegated authority names, with the test results attached.
- 06
Release and measure
A test group often goes first. The Business Outcomes Layer shows what the change did to conversion, revenue, cost, losses and collections, against plan, so each change becomes evidence for the next.
Three tests for every change
A change passes only if it improves one of these without harming the others beyond your risk appetite:
Conversion and volume
More customers approved, accepting, staying and taking what suits them.
Cost
A lower cost to acquire, serve and collect.
Credit quality and fair outcomes
Losses within appetite, and customers treated fairly at each stage.
Collections that recover what is owed, and leave customers able to keep paying.
Models and strategies help decide who to contact, when, through which channel and by whom, and which arrangement is likely to hold. We build them with your collections, risk and conduct teams, and run them in Zoral Decision Engine alongside the collections systems, diallers and agencies you already use.
Who will pay, and who needs help
Propensity-to-pay models separate customers who will pay after a reminder from those who need a conversation or support.
Reaching the right person
Right-party-contact models suggest the numbers, channels and times most likely to reach the customer, within your contact rules and the customer’s consent and preferences.
Queues in the right order
Work ordered by expected recovery, risk and the customer’s known circumstances, not by balance alone.
Timing and channel
Reminders, messages and calls are sequenced from before the due date. Contact frequency stays within your policy.
The right collector
Customers matched with the collectors most likely to agree a workable arrangement, integrated with your existing dialler.
Failed payments, foreseen
Payments likely to fail are predicted, so a reminder, a new payment date or support can be offered in time.
Arrangements customers can keep
Payment plans are set against what each customer can afford. Refinancing is part of a responsible strategy where it lowers the burden, with the total cost made clear.
Measured on the figures your committee uses. Collected amount, recovery rate, cost to collect, contact rate and arrangement rate, alongside arrangements kept and how customers in difficulty fared, each compared with your current strategy.
Our collections experience
Our collections models have run in consumer lending portfolios and alongside specialist collections firms with experienced analytics teams of their own. Our segmentation combines the value an account is likely to return with the chance of reaching the right person. Tested on the same accounts with the same effort, it concentrated more of the cash collected into fewer accounts. Collectors could then reach the customers most likely to pay, and give more time to those who needed support, without the team growing. Each engagement starts with a check of the data behind collections and follows the same data-science-driven MVP methodology as our credit work.
Fair treatment is designed into the strategy, not checked after it.
Pressing harder on customers who cannot pay builds up conduct risk and losses, so safeguards are built into the strategy:
Customers who may be vulnerable are routed to trained staff and taken out of automated contact where your policy says so.
Forbearance, such as reduced payments or a restructured term, is part of the strategy, not an exception a persistent customer has to find.
Contact frequency, times and channels follow each market’s rules and the customer’s stated preferences.
Offers and limit increases leave out customers showing signs of financial difficulty.
Outcomes are compared across customer groups, so differences appear in your own reports first.
We design for what supervisors expect, under the UK’s Consumer Duty and comparable rules elsewhere. Your obligations remain yours. Zoral gives you the controls and evidence to meet them.
Every recalibration goes through your approvals.
Optimisation should not mean models that shift your risk profile on their own. Each change to a model, policy, price or collections strategy arrives with its test results and is approved by the people your delegated authority names. Models pass your model risk process and run in shadow first. Each change is versioned and reversible. The record shows who proposed and approved it.
AI agents, configured to your policies and workflows, help analysts prepare scenarios and collectors prepare for a call. They are guardrailed, work under human-in-the-loop supervision, act only through the permissions and approvals you set, and what they do is recorded. Agents do not make credit decisions or decide how a customer in arrears is treated. Those follow rules and models your bank approves, and accountability stays with your bank.
Implemented on Zoral fOS, or applied in the systems you run.
Atelier implements Zoral fOS and its solutions, so the people who design a strategy also configure it. Models and strategies run in Zoral Decision Engine in real time. Zoral Data Fabric records each decision and outcome, and the Business Outcomes Layer measures each change against plan. Your teams then make further changes themselves, under the same governance. Some engagements use no Zoral product, such as a collections review applied in the systems you run today. We work directly or alongside your advisers, integrators and agencies.
Related solutions: Retail Banking · Credit Cards · BNPL & Point-of-Sale Finance
Related Atelier services: Portfolio optimisation and management · Credit risk and decisioning · Data, AI & Automation
Ways to start.
Lifecycle value review
For one product, where value is gained and lost from acquisition to recovery, measured on your data, ending with changes ranked by value.
Collections strategy review
Segmentation, contact strategy, arrangements and agency use, reviewed with your collections and conduct teams, with practical recommendations.
One portfolio, tuned against plan
We set the plan and baseline on Zoral fOS, then take the first changes through testing, approval and release with your team.
Models for one stage
Propensity-to-pay, right-party-contact, retention or lifetime-value models, built with Zoral’s data-science-driven MVP methodology and shadow-tested against your current ones.
Frequently asked questions
Customers and conduct
Will optimisation push customers into products they don’t need?
It is built not to. Offers go only to customers who pass your suitability and affordability checks and whose preferences allow contact. Customers in financial difficulty are left out. Results are judged on later repayment and customer outcomes, not take-up alone. You decide which offers are made, and to whom.
How do you treat customers in difficulty?
Early, and as people rather than balances. Strategies offer the forbearance your policy allows, route customers who may be vulnerable to trained staff, and set arrangements against what the customer can afford. Each contact and arrangement is recorded, so you can show how they were treated.
Collections operations
Can you work with our existing collections systems and agencies?
Yes. Zoral connects to collections systems, diallers and payment systems through documented interfaces. Strategies can be configured to allocate accounts between your teams and agencies, and agency results compared on the same measures. You decide what changes, and when.
Do models replace our collectors’ judgement?
No. Models suggest priority, timing, channel and arrangement. Collectors can depart from a suggestion, with the reason recorded, and those reasons help tune the next strategy.
Control and results
How do changes get approved?
Through the authorities you already have. A change is tested on past data and approved by the person or committee your delegated authority names. It can start on a test group, and can be reversed with its record intact. Model changes also pass your model risk process.
How do we know it’s working?
Each change is measured against your plan and against the strategy it replaced, by product, channel and segment. Where results drift, the Business Outcomes Layer shows where the gap comes from. Results depend on your products, data and starting point, so we measure against your own baseline.
Founded in 2004, Zoral has delivered over 100 projects, and our people bring 350+ years of combined experience. We have worked with clients regulated by the FCA in the UK, the CFPB in the US and the central banks of Malaysia and Thailand. References are available in confidence.
Start with the stage where value slips away.
Bring us one product: the conversion you want to raise, the customers you lose at renewal, or the arrears that keep rising. A senior Atelier team with credit, collections and data science experience will look at where value is lost, what your data can support, and which change to test first.
Contact Us
Prefer to pick a time? Schedule a meeting
Automate lending, banking and finance operations.
London
+44 20 3963 2002Miami
We are on Socials