The product owner told me it was the first software project he'd ever seen delivered on time.
I'd taken ownership of the team and the work, built direct relationships with the stakeholders, and stayed close enough to them to see expectations drifting before they became problems. None of it was especially sophisticated, but it worked, and I was proud of it.
Then I noticed something. He'd joined four months before me. He'd hired me, and the project was his responsibility. But as the work went on, the people who mattered increasingly associated the delivery with me rather than with him. Someone he'd brought in was becoming the main contact for his key stakeholders, and I could see what that was doing to his position.
So I raised it with him. I said I'd noticed the tension and wanted to work out how we could make it work for both of us.
The conversation didn't go anywhere. It was acknowledged and moved past, and I took that as an answer.
What I did instead
I pulled back without saying so. I stopped going straight to stakeholders, routed things through him, and held back in conversations I'd previously led.
Nothing collapsed, which made it easy to pretend the trade-off was harmless. But the thing that had made the project work went with it. Early warnings stopped reaching me, and the conversations that had kept everyone honest had nowhere to go.
What I'd do now
I spotted the problem early and I named it, and I'd do both again. Where I went wrong was treating a conversation that hadn't landed as if it had been settled. When something important is at stake and the conversation stalls, the job is to go back and say: "We haven't sorted this, and we need to." If you don't, the conversation still happens, later, in worse conditions and with fewer options.
I don't know whether pushing harder would have changed the outcome. I do know that I now treat an ambiguous answer as the start of the conversation.