Key Takeaways
- Agents like Transformance’s Vero work from each customer’s payment history, promises and disputes, and reminders pause when a dispute or promise to pay is logged.
- No published study measures B2B churn after dunning; a reminder letter raised payment by about 25 percentage points in an Australian tax experiment (Gillitzer and Sinning, IZA discussion paper, 2018).
- More than half of surveyed Western European businesses name liquidity pressure as the main reason customers pay late (Atradius, 2026), so a first reminder should assume an oversight.
- Pause on a dispute, a payment in transit, a dated promise, a partial payment or a changed contact; restart on a stated date or event.
- A person takes over at a dispute, a broken promise, a strategic account, a large balance and before any formal step.
In This Article
- Key Takeaways
- Does automated dunning damage B2B customer relationships?
- What does the evidence show, and what is only convention?
- What should stop an automated reminder?
- How do you write a friendly payment reminder that still gets paid?
- Which channel suits which touch?
- When should a person take over from the automation?
- How do you know whether dunning is harming relationships?
- How does Transformance automate dunning without losing the relationship?
Does automated dunning damage B2B customer relationships?
Not by being automated: dunning hurts when it chases through a dispute or a promise, or treats key accounts like the long tail; agents like Transformance’s Vero avoid both.
Tone, channel and handover by customer segment (net 30; a starting policy, not a benchmark)
Segment names match our B2B dunning cadence by segment, which sets the days; this page sets tone, channel and handover.
No published study measures whether automated dunning raises B2B churn. What is measured: a reminder letter raised the probability of payment by about 25 percentage points among Australian business taxpayers (Gillitzer and Sinning, IZA discussion paper, 2018), and reminders sent after a promise to pay could trigger a repeat default among 32,000 Latvian borrowers in default (Saulitis, 2023). Both are tax or consumer debt, not trade invoices.
Automated dunning is the sending of payment reminders and notices by rule instead of by hand, with stage, text and timing set in advance.
What goes wrong is a generic reminder after a dispute, a payment or a promise. The matrix is our policy, built on the evidence below and common practice; test it per segment.
Takeaway: automation repeats whatever policy you give it, so write the policy for context, not for averages.
What does the evidence show, and what is only convention?
Only a few things are measured: reminders work, small timing differences matter little and the effect of personalization is mixed; the friendly-to-firm ladder and every day count are convention.
The first touch should assume an oversight: more than half of surveyed Western European businesses name liquidity pressure as the main reason B2B customers pay late, and administrative bottlenecks, banking delays and e-invoicing issues add delays even where cash is available (Atradius, May 2026).
Not measured: churn or complaint rates after dunning, a best tone for trade invoices, a best day count. All the experiments above concern tax or consumer debt. “Friendly, then firm, then formal” is sensible convention; nobody has shown it is optimal, and the evidence on tone is mixed.
Takeaway: the evidence backs reminding early and pausing after a promise; everything else is a choice to test.
What should stop an automated reminder?
Pause an item when it is disputed, paid or in transit, under a dated promise, partly paid, or its contact has changed; restart on a stated date or event.
Order of work matters most: cash application runs before the dunning run, so a customer who has paid is never reminded (partial payments and deductions in AR).
Disputes follow a separate track with one owner and a deadline (B2B invoice dispute workflow; AI dispute management). The promise row has direct evidence: in the Latvian experiments, reminders after a promise could trigger a repeat default (consumer debt, not trade invoices; how AI tracks promise to pay).
A reminder that lands in spam hurts too (why dunning emails land in spam).
Takeaway: every pause needs an owner and a restart date, or the invoice disappears.
How do you write a friendly payment reminder that still gets paid?
Name the invoice, amount and due date, assume an oversight, give one clear action and sign with a person; raise firmness by stage, not by capital letters.
These are opening lines; six full emails are in our dunning email templates.
Rules from practice, not measured results:
- Assume oversight until the customer says otherwise.
- One action per message: a date, a confirmation or a payment.
- A named sender and a reply-to address a person reads.
- No all-caps subject lines.
- Quote the customer’s own promise when one exists.
Wording effects in the experiments are small and mixed: the ladder is service design, not a proven optimum. A friendly payment reminder works because it is accurate about why most invoices are late.
Takeaway: firmness comes from facts and dates, never from volume or formatting.
Which channel suits which touch?
Use email for the early touches if it arrives, the phone from the human-contact stage, the customer’s AP portal to check status, and a written notice for the formal record.
- Email. Cheap and documentable; right for the routine.
- Phone. In the Colombian trial, calls worked when paired with an easy, immediate way to act. What to say: AR collections call script by dunning stage.
- AP portal. Check invoice status before you chase; many “late” invoices are stuck in approval (AP portals).
- Letter or formal email. A formal notice should be provable, so keep it in writing.
Channel follows segment too: call a strategic account before any formal letter, not a long-tail account on day 7 (collecting from thousands of small B2B accounts).
Contact people in their own language. Transformance’s AI calling agent makes 15-20 calls an hour in 70+ languages, against 15-20 a day for a human collector (Transformance platform data, 2026), and says at the start of each call that it is an AI (AI calling agent guide).
Takeaway: the channel tells the customer how seriously you take them; match it to segment and stage.
When should a person take over from the automation?
A person takes over at a dispute, a broken promise, a strategic account, a large balance, any reply that is not a payment date, and before any formal step.
- Dispute: the dispute owner.
- Broken promise: the collector.
- Strategic account or open renewal: the account owner.
- Balance above your threshold: a senior collector.
- Angry reply or a service problem: the account owner and sales.
- Supply hold, agency referral or legal step: a named decision-maker under your policy.
Set thresholds in an escalation matrix; groups keep one baseline with entity overrides (dunning across entities).
Most handovers start with a reply. A shared AR inbox mixes remittances, disputes, promises and questions; triage by intent decides which replies a system may answer (how AI triages a shared AR inbox).
What to ask any dunning tool:
- Do items pause on dispute, payment and promise?
- Can segments differ in tone, channel and timing?
- Who approves a formal notice?
- Can a person take one item over?
- Does the log show what was sent and why?
- Does it work with your ERP?
Takeaway: the handover is a rule you write down, with an owner per trigger.
How do you know whether dunning is harming relationships?
Watch leading indicators you already hold: reminders sent to paid or disputed items, disputes after a reminder, complaints, sales escalations, days to pay and promises kept, per segment.
No benchmark values exist for these rates; track your own (AR collections KPI guide). Silence is its own pattern (when a B2B customer does not respond); ranking is in how to prioritize AR collections with AI.
Our method: test two tones inside one segment for eight weeks, split by odd and even customer number; never compare across segments.
Takeaway: fix reminders to paid or disputed items first; that number should be zero.
How does Transformance automate dunning without losing the relationship?
Transformance’s CollectPulse pauses reminders when a dispute or promise to pay is logged, sets tone, channel and timing per segment, and Vero works from each customer’s full history.
Rules and pauses. In CollectPulse, tone, channel and timing are set per customer tier or segment, so the strategic and long-tail rows of the matrix run side by side. Reminders pause automatically when a dispute or a promise to pay is logged.
Context. Vero, Transformance’s AI agent, works on Customer 360: one graph per customer with payment history, promises to pay, disputes and deductions. Vero investigates exceptions from it plus internal and external sources and proposes the next step with the evidence; a person approves.
Calls. The AI calling agent agrees payment plans within rules you set, sends a written confirmation after the call (promise to pay or plan) and records each call in the audit trail.
Replies. Vero classifies emails in the shared AR inbox by intent (remittance, dispute, promise to pay, query), drafts replies and sends routine ones, such as invoice copies, within rules you set. The rest goes to a person; a bank-detail change, a credit or refund request, a write-off or legal wording always does.
Results. CollectPulse actions 100% of overdue invoices within 24 hours, against 30-40% for manual teams; Vero handles 60-80% of routine touches; customers capture 3x more promises to pay than with manual work; DSO falls 8-12 days on average (Transformance customer data, 2026).
Control. Multi-level approvals set thresholds by amount, entity and customer (who approves what); the audit trail records what the system did, what the user did and who approved.
ERPs and time. SAP S/4HANA, SAP ECC, NetSuite, Microsoft Dynamics 365 and Odoo run in production with customers; Oracle and Infor are supported. The ERP connection typically takes 2-3 days, your IT team may be involved, and full rollout takes up to 4-8 weeks (Transformance customer data, 2026). See ERP integrations.
Takeaway: the routine runs by rule per segment; the judgment, with Vero’s evidence, stays with your team.
Who Transformance is for
- ERP: finance teams that run AR in SAP S/4HANA (on-premises, private or public cloud), SAP ECC, NetSuite, Microsoft Dynamics 365 Finance & Operations or Business Central, or Odoo, all running in production with customers; Oracle and Infor are supported.
- Size: mid-market to large enterprise, 250 to 2,000+ employees, including groups with several entities and shared service centers.
- Payment volume: from under 1,000 to over 100,000 payments a month.
- Industries: manufacturing, consumer goods and food and beverage, logistics, e-commerce, agriculture, energy and professional services.
- Scope: cash application alone, or the full order-to-cash: cash application, deductions, AR collections and cash forecasting.
Conclusion: automate the routine, write down the judgment
Automated dunning damages relationships when it is blind to context. Pause on disputes and promises, set tone and channel per segment, name who takes over, and measure harm with your own data. Book a Call.
Key terms
- Dunning: the structured sequence of reminders and notices for overdue invoices, with rising firmness.
- Promise to pay: a dated commitment from the customer; it pauses reminders until that date.
- Dispute management: handling contested invoices outside the standard cadence.
- Escalation matrix: who takes over at which stage, and with what authority.
- Delinquency bucket: an aging band used to assign the next action.
- Past due: an invoice not paid by its due date.
Related guides
- Which ERP systems does Transformance integrate with?
- What B2B dunning cadence works by customer segment and terms?
- What should an AR collections call say at each dunning stage?
- What does a B2B invoice dispute workflow look like, step by step?
- How can AI triage a shared accounts receivable inbox?
- Dunning process: email templates and automation guide
Frequently Asked Questions
Does automated dunning damage B2B customer relationships?
Not by being automated: it damages them when it chases through a dispute, a payment in transit or a promise, or treats strategic accounts like the long tail. No published study measures churn after dunning.
Should the first payment reminder be friendly?
Yes. More than half of surveyed Western European businesses name liquidity pressure as the main reason customers pay late (Atradius, 2026), so a first touch that assumes an oversight is usually accurate.
When should a person take over from automated reminders?
At a dispute, a broken promise, a strategic account, a balance above your threshold, any reply that is not a payment date, and before a supply hold or agency referral.
Which ERP systems does Transformance work with?
SAP S/4HANA (on-premises, private or public cloud), SAP ECC, NetSuite, Microsoft Dynamics 365 Finance & Operations and Business Central, and Odoo, all running in production with customers; Oracle and Infor are supported. The ERP connection typically takes 2-3 days (Transformance customer data, 2026); your IT team may be involved.
How large does a company need to be to automate dunning this way?
Transformance is built for mid-market to large enterprise teams, 250 to 2,000+ employees, handling from under 1,000 to over 100,000 payments a month.
How long does it take, and what does it cost?
A full rollout takes up to 4-8 weeks (Transformance customer data, 2026). Cost depends on scope (cash application alone or the full order-to-cash) and the number of entities: Book a Call.
How can you tell whether dunning is harming customer relationships?
Track reminders sent to paid or disputed items (should be zero), disputes within a week of a reminder, complaints, sales escalations, days to pay and promises kept per segment; then test two tones inside one segment for eight weeks.
SAP, SAP S/4HANA, Microsoft, Dynamics 365, Business Central, Oracle, NetSuite, Odoo and Infor are trademarks of their respective owners (SAP SE, Microsoft Corporation, Oracle Corporation, Odoo S.A., Infor). Transformance is not affiliated with or endorsed by them.




