Philippines staffing blog ·
Keep a support promise when a dependent team slips
A workflow for protecting a customer promise when an engineering, billing, or vendor dependency misses its date.

Direct answer
Operational guidance
Why a promise survives a dependency
A support promise belongs to the customer even when the work depends on another team. The customer agreed to a date, a channel, and an outcome with the desk, not with the engineering queue behind it. When the dependency slips, the desk still owns the next update. Pretending the promise moved with the work damages trust faster than the delay itself. The rule is simple: the date may change, but the ownership of the message does not.
Teams often wait for a confirmed new date before telling the customer anything. That instinct protects the desk from saying something wrong, but it leaves the customer with silence. A useful update can say what is known, what changed, and when the next check will happen without inventing a date. The customer needs a reliable rhythm more than a perfect prediction. Silence is the only update that is always wrong.
Separate the promise from the plan
Write the customer promise as one sentence with a date and an outcome. Write the internal plan as the sequence of dependencies that must finish first. Keep the two in separate fields so a change in the plan does not silently rewrite the promise. When a dependency slips, compare the new plan with the promise before choosing a response. That comparison tells the desk whether it needs a small update or a new agreement.
The promise should include what happens if the date cannot be met. A customer may accept a later date, a partial delivery, or a refund path. Record which of those options the customer approved at the start. Then the desk can act without reopening the whole negotiation during a crisis. An agreed fallback is a promise too, and it deserves the same respect.
Send an update before the deadline
Send the first update as soon as the slip is credible, not after the deadline has passed. Say that the original date is at risk and explain the reason in plain language. Name the next checkpoint and the person who will report at that time. Do not blame another team by name or imply that the delay is nobody's responsibility. The customer cares about the outcome and the next reliable contact.
Use the channel the customer chose and copy the update into the ticket. If the customer asked for a call, make the call even when the news is incomplete. A short call with a clear next step is better than an email that arrives after the promised hour. Record what was said and what the customer expects next. The next owner should not have to reconstruct the conversation from memory.
Name the new date and its basis
When a new date is offered, explain what it depends on and how confident the desk is. A date based on a completed dependency is stronger than a date based on a team's estimate. Label the difference so the customer can judge the risk. Offer a checkpoint before the new date rather than waiting to see whether it holds. That checkpoint is the promise the desk can actually keep.
Avoid moving a date more than once without a written change to the plan. Repeated slips usually mean the dependency is not understood or not owned. Escalate the pattern instead of sending another apology. A manager may need to change scope, add capacity, or renegotiate the outcome. The customer should hear that decision from the desk, not discover it by watching the date move again.
A worked example
A customer is promised a data export by Friday and the engineering team reports on Wednesday that a permission fix will take another week. The support owner confirms the dependency, checks the customer's fallback agreement, and calls the customer before the original deadline. The call explains that the full export cannot be ready, offers a partial export of the available records, and sets a check-in for Monday. The customer accepts the partial file and keeps the full export as the next milestone.
The agent sends the partial export with a short note that names what is included and what is still pending. On Monday the engineering owner confirms the permission fix and the desk sends the revised date with its basis. The customer receives one thread with the original promise, the change, and the recovery. No one claims that the delay was unavoidable or that the new date is guaranteed. The record shows an honest sequence rather than a silent gap.
What the manager decides
The manager decides which dependencies count as customer-facing promises and which remain internal estimates. A promise that appears in a contract or a public status page needs a stricter update rule. Name the owner who may change a customer date and the evidence required before the change. Publish the escalation path for a dependency that misses two checkpoints. A clear path keeps agents from bargaining alone with an angry customer.
The manager also decides how much detail the customer should receive. Too little detail feels evasive, while too much internal process shifts the burden to the customer. Give the reason at the level that helps the customer plan, not the full incident history. Review a sample of delay updates for clarity and honesty. The best update leaves the customer knowing exactly what happens next.
Train the update habit
Use a real slip from the past month and ask the team to write the update they would have wanted to receive. Compare it with what the customer actually got. Look for missing dates, vague reasons, and unclear ownership. Rewrite one message together and save it as a pattern for the next dependency delay. The exercise should produce language, not blame.
Add delay updates to the quality sample and check that each one names a checkpoint and an owner. Look for tickets where the customer asked for an update before the desk sent one. That is a signal that the internal warning arrived too late. Coach the team to raise a risk when it is first credible. Early honesty is cheaper than a late explanation.
Close and review
Close the promise only when the outcome is delivered or the customer accepts a documented alternative. Record which condition was met and who confirmed it. If the dependency is still open, keep the ticket assigned and the next checkpoint visible. Do not close a case simply because the internal team has stopped replying. The customer's promise is not complete until the customer has the result.
Review dependency delays each month and group them by the team or system that caused them. A repeated cause deserves a process fix rather than a better apology template. Update the promise language when customers consistently misunderstand the fallback. Keep the review short and the actions named. The goal is fewer surprises, not fewer promises.
Authoritative sources
- ISO 9001:2015 quality management systems
- Atlassian incident management handbook
- ISO 18295-1:2017 customer contact centres