1) Power factor has nothing to do with power based penalties. If your company is being penalized on power factor, that is real and solvable, but not related to how your facility is started up. Add PFC caps to the motors, starting with the largest ones, until your facility PF is .95 or better, after that, don't waste any more money. That's the typical penalty point.
2) Unless you can somehow stretch the starting time out to more than 30 minutes, adding soft starters will likely have zero effect (directly) on what are called "peak Demand" charges, if that's what he is thinking. It's one of those "alternative facts" that is repeated so many times by people marketing soft starters that it begins to sound true, but from a pure engineering fact based standpoint, it's bogus. 99.000% of power utilities bill for Peak Demand based on what is called a "demand window", a snapshot of a 15 to 30 minute window of time in which the average power for that period is highest. Your utility likely publishes their Peak Demand schedule that will lay out that demand window length. But bottom line, the 20-30 second surge you might see from starting motors across-the-line is actually insignificant, and don't be swayed by people who tell you otherwise. That myth is as old as the hills and many many people have a vested interest in maintaining it, but if you still doubt me, call the utility and ask them how much your bill will drop by adding soft starters. They will tell you zero.
There are plenty of other REAL benefits to using soft starters, I happen to really like them, but the reduction of Peak Demand charges is not one of them. The reason the marketing people get away with this alternative fact is that in maybe a small handful (as in less than a dozen across the US) small utilities, mostly that buy power from other larger utilities, they did at one time use what were called "ratcheting instantaneous demand meters". I have not seen one of those in over 20 years now, yet the myth persists.
If you want to help your company avoid penalties the best strategy, and this by the way DOES often involve a controls engineer, is to thoroughly and carefully evaluate the real need to run any and all machines, then adopt control strategies that eliminate wasteful practices. I'll give you an example.
I did work at a wood mill-works, where they make molded wood products for furniture, picture frames, mantles, trim etc. They had dozens of machines like molders, saws, routers, chippers etc., each once had it's own dust collection blower that ran whenever the machine ran, plus many used compressed air and hydraulics, so any running machine was in fact running MULTIPLE machines. The first thing I implemented was that in observing the 2 shift operations for a week, I found that each machine was turned on by the first shift and turned off at the end of the second shift; NEVER off during breaks, lunch, dinner etc. Even if the machine was not being used, it was left running. When I asked operators why, their response was that they were told that "it cost more to re-start a machine than it does to leave it running for an hour, let alone for a 15 minute break." This of course was based somewhat on that same myth about peak demand charges, but also on a real issue of larger machine start-up sometimes causing a voltage drop that made the HID lights re-strike, which meant it was dark for 10 minutes. So we implemented soft starters on every machine over 20HP, mostly just to change their behavior, but also to prevent that voltage drop. That ALONE dropped their demand charges, not because the peak starting current was reduced, but because the OVERALL load demand was reduced and less energy was wasted.
The second phase was that we tied all of the dust collectors together with one common air duct system, then evaluated how much suction was required at each machine. When any given combination of machines were running, the dust collection control system decided which of the multiple dust collectors to turn on in a combination that got the desired effect. This of course could have been easier with a few collectors and VFDs, but they already had all of these multiple smaller collectors installed, so we used them. Overall this then reduced the overall energy consumption another 18% per year. The modifications, PLC and software paid for itself in under 60 days.
The point is, your boss has given you a golden opportunity to be a hero here. Use it, just not in the way he thinks it needs to be done. So my last piece of advice; don't fight him, manipulate him slowly into thinking it was his idea. Ask first for an opportunity to do a thorough study of energy usage so that you can make optimum suggestions without wasting money, then use that information you gather to come up with solutions that, as I said, usually fall into the controls realm anyway.
Good luck!