A practical response to missed commitments that separates causes, contains exposure, and makes restoration conditional on evidence.

Robert Kirk, B2B Analytics Insights
An invoice becomes overdue. The account owner hears a reasonable explanation, collections hears another promise, and Procurement worries about interrupting an important relationship. Someone says the customer has broken trust. Before that judgment drives the next decision, the team needs to identify what actually broke.
Was it the commitment, the process, the financial capacity, or our interpretation of the evidence? More than one can be involved. Repair begins by separating them, while protecting the cash flow already exposed. Waiting for certainty is not a reason to leave further exposure uncontrolled.
Repair protects the next cooperative outcome
The Prisoner’s Dilemma gives a precise reason to consider repair. In Axelrod’s illustrative rewards, mutual cooperation earns 3 for each player per round, whereas mutual defection earns 1 each. Breaking a cycle of mutual noncooperation can therefore restore value for both players. Those numbers describe a game, not a predicted financial recovery. Axelrod, authorized adaptation
Under one explicitly simplified assumption, a defection ends cooperation permanently. If δ is the effective future weight, continued cooperation is worth 3 / (1 − δ) and defecting now is worth 5 + δ / (1 − δ). Cooperation is at least as valuable at δ ≥ 0.5 under that reward structure and response. This is an illustrative calculation, not the threshold for Tit for Tat or a prescription to punish a customer permanently.
The business implication is to examine the value of the next dependable exchange, not merely the anger created by the last failed one. The financial qualification is equally important: the prospect of another profitable transaction does not establish capacity to pay. Repair needs both a credible relationship path and a supportable exposure decision.
Define the breach before defining the response
Write down the agreed obligation, the observed departure, and the evidence supporting that observation. Distinguish a missed payment from a disputed invoice and a missed delivery from an unconfirmed receipt. Record what remains unknown. This creates a shared starting point without pretending that everyone already agrees about the cause.
Keep identity, authority, financial capacity, and willingness separate. A legitimate business may face financial strain; an authorized person may make a promise the business cannot keep. A persuasive explanation does not verify identity or establish capacity. An overdue balance alone does not establish deliberate deception.
I would revisit the same five credit questions, rather than replacing analysis with a label about character. Their sequence keeps the decision connected to permission, recovery, timing, exposure, and change. Each answer should identify its evidence and its unresolved uncertainty.
- Is this a counterparty I trust and am allowed to do business with?
- Will I ever get paid back?
- Will I be paid back on time?
- How much should I extend?
- What has changed?
Apply the Framework to the next decision
My proposed Framework is Verify, Cooperate within bounds, Observe, Respond, Repair or Exit. It is a practical synthesis, not an established result from cooperation research. In a strained relationship, the steps become a disciplined way to decide what can continue and what must change.
Verify
Confirm the entity, the authority behind the latest request, and the underlying transaction records. Compare competing explanations with relevant documents and contact the appropriate person through an established channel. Review our own contribution to the problem. A wrong invoice or unrecorded acceptance should not become evidence against the other business.
Cooperate within bounds
Decide what activity can safely continue while the issue is investigated. Credit should own the exposure decision, with Procurement and operations contributing continuity requirements. State the temporary boundary and its review condition. Do not let commercial urgency create an exception that nobody explicitly approved.
Observe
Specify what would distinguish the explanations under consideration. A confirmed invoice correction supports a process explanation; a missed revised commitment calls for further review. Relevant external financial information can inform capacity, but may be incomplete or delayed. Keep observations separate from conclusions, and record the source and timing of each.
Respond
Choose a proportionate action that addresses the likely cause and the potential loss. Correct a process failure, reassess terms when capacity changes, and escalate unresolved identity or authority concerns. Explain the action without claiming more than the evidence supports. Include a route for the counterparty to challenge inaccurate information.
Repair or Exit
Write the restoration conditions before lifting restrictions. Name the evidence required, who can approve the change, and what continued failure would trigger. Repair should depend on observable commitments and acceptable exposure, not an apology alone. Exit may be the responsible outcome when identity remains unresolved, capacity is insufficient, or credible commitments cannot be established.
Illustrative worked example: separate the explanations
Consider a hypothetical distributor waiting for payment from a customer that also provides an important service. This is an illustrative scenario, not a reported engagement or result. The customer says payment is delayed because the distributor’s invoice lacks a required reference. The account team initially interprets the delay as unwillingness to pay.
The reviewer tests three competing explanations: an invoice problem, financial strain, and intentional delay. Internal records confirm the missing reference, so the distributor corrects its own error and records that finding. The customer confirms receipt through an established contact. The reviewer does not infer financial strength from this exchange; capacity still requires separate assessment.
Credit defines a temporary boundary for new exposure while Procurement reviews service continuity. The parties document the next payment commitment and the evidence needed to confirm performance. The reviewer sets a decision point linked to that commitment. If performance occurs and the capacity evidence supports it, the authorized owner can reconsider the restriction.
If the revised commitment is missed, the team updates its assessment rather than repeating the original explanation. It may seek stronger evidence, revise terms, or decide not to extend further credit. The invoice error remains our responsibility even if another concern later emerges. Accountability is not something either party earns only when the other behaves perfectly.
Reopen the relevant ten phases before restoring terms
Relationship repair is not a separate approval process. It should reopen the relevant parts of the 10 Phase Credit Checking Framework, with the original decision as the comparison point.
For a payment dispute, revisit phase 6 to distinguish a specific administrative obstacle from a broader payment pattern. Revisit phase 4 when cash generation or liquidity may have changed, and phase 5 when funding or competing obligations affect repayment assumptions. Review phase 3 if performance has deteriorated. A changed contact or instruction can also reopen phases 1 and 2; a restored conversation does not resolve identity or authority.
Use phases 7 and 8 to test relevant new information and owner or related entity dependencies, without treating an unverified signal as a conclusion. Phase 9 then specifies the next review trigger and its owner. Phase 10 records the revised limit, terms, additional protection if justified, and the uncertainty we knowingly retain.
Consider the portfolio as well as this account. Restoring several related customers at once can increase dependence on the same underlying cash source or market. A credible individual recovery plan should not silently override aggregate exposure limits. The relationship decision and the concentration decision need to agree.
The practical distinction is simple: correct our error because it is our responsibility; restore commercial terms only after the relevant financial and control conditions support doing so. Kindness governs how we respond. Evidence and risk appetite govern what we extend.
Forgiveness does not reset the exposure
Axelrod’s later work describes how misunderstanding and implementation errors can cause simple reciprocity to amplify conflict. Generosity and contrition can help address those problems in the modeled settings. He also makes clear that there is no universally best strategy. The useful business lesson is to examine error and response, not mechanically copy Tit for Tat. Axelrod, Launching The Evolution of Cooperation
I distinguish forgiveness from restoration of commercial terms. We can stop treating a past failure as a permanent judgment while still requiring evidence before accepting new exposure. Equally, when our mistake caused the dispute, we should correct it and reconsider any restriction based on that mistake. Contrition must change our decision process, not just our language.
Preserve future value without advancing it on faith
The mathematical example assumes players can deliver the cooperative action. In credit management, that assumption must be tested. A customer who lacks liquidity cannot settle an invoice merely by deciding to cooperate. Reopening phases 4 through 6 separates capacity, funding constraints, and payment performance from our interpretation of intent.
The example also assumes that behavior is observed accurately. Our own invoice error shows why that can fail. Phase 9 should connect a verified event to an accountable review rather than convert a disputed signal into automatic retaliation. Phase 10 should document what exposure may continue and what evidence permits restoration.
This is the joint lesson: the value of future cooperation is a reason to investigate a workable repair, while the ten phases determine whether its financial conditions are credible. We can be willing to repair a relationship without being willing to extend the same terms.
Make repair reciprocal and reviewable
AI can help assemble a chronology, identify document inconsistencies, and prepare alternative explanations for review. Analytical models should inform capacity and payment risk before narrative interpretation takes over. A named human must approve material changes, assess contested evidence, and own any exit decision. The record should explain both the action and the conditions for reconsidering it.
For Procurement and finance leaders, the goal is dependable performance and cash flow that both businesses can plan around. That aim must also respect customers, employees, and others affected by the arrangement. Cooperation is not a public good simply because both parties benefit. Repair should restore legitimate commitments, not conceal a problem or transfer its cost to someone outside the conversation.
At the next review, ask what changed in the evidence and whether we kept our own commitments. Restore only the confidence the evidence supports. That is how a broken relationship can become a better governed one, or end without confusion about why.
Internal Knowledge + External Insights = Competitive Intelligence
