A strong answer. “First I'd check whether the delay is capacity or priority, because they look identical and need opposite responses. If their integration owner has been pulled onto something else, this is a priority problem and only their sponsor can fix it. If they genuinely cannot do it, that's a capacity problem and the answer might be that we do it, scoped and possibly charged.

Then I'd re-baseline rather than compress. Trying to hold the original date by shortening testing is how implementations go live badly, and a bad go-live costs more than a late one — it's the thing that turns a customer into a churn risk in year one.

What I'd bring to the sponsor conversation is not the complaint but the options: go live on the original date with reduced scope, move the date by three weeks, or we take on the integration work. Each with what it costs them. Sponsors respond to choices far better than to status.

And I'd have flagged this in week one rather than week four. Every implementation I've seen go badly went badly because the dependency was visible early and nobody named it while it was still cheap.”