I would stop treating 15 posts a week as the goal and use the saved time to learn what the audience will actually respond to. For the next month, try 3 to 5 stronger posts a week, with each one built around one specific buyer problem rather than a general company message.
Before changing everything, pull the last 30 posts into a simple table with the topic, format, opening line, reach, watch time, saves, shares, profile visits, and replies. Look for patterns in the middle of the distribution, not just the one post with the biggest spike. That usually gives you a more honest signal than likes alone.
I’d also agree with your boss on one definition of “real engagement” before the test starts. For a B2B account, qualified replies, conversations with the right roles, and profile visits may matter much more than broad likes, and that makes the content decisions easier to defend.