This feels like an advertisement for agentic AI. Though I agree wholeheartedly that teams can be proactive about blockers, this is only introducing more dependency that can lead to wasted effort or costs. The primary proposed solution is basically accepting delay by temporarily accepting AI substitutes for blocking teams and processes and hoping that filling the AI requirements will fill the actual team's requirements when the blocker is actually dealt with. Basically option 2 or 3 with the trust that adding another dependency on AI will magically fix things. When it doesn't, it will be wasted time meeting a hallucinated response. Not to mention that when the cost of tokens starts going up to meet the market predictions of these technologies, this is going to constitute a significant financial cost for organizations and teams.
I think about this as the traffic jam conundrum. You can wait in the line of cars idling or take the exit to local roads, navigating traffic lights and driving extra miles to stay in motion while rejoining the now unjammed traffic further on right behind the same car. I'll admit I do this, but I won't pretend that the benefit after consuming additional gas was anything more than avoiding waiting for the sake of my nerves.