Influencing without authority

My job runs almost entirely on influence without authority. I work on AI across business units at One Planet Group, and none of those units report to me. Every leader I work with owns a P&L and already had plans for the year before I showed up wanting to change how their team works, and nobody is required to say yes.

I underrated storytelling for years. A business unit leader hears pitches all week, and what actually holds their attention is a story where they can see their own operation in it. So before one of these meetings I write out what a normal day looks like for their team right now, where the time and revenue go that they wish it didn't, and what that same day could look like six months after a pilot. Shreyas Doshi has written about this, and he puts storytelling near the front of where influence comes from, well ahead of title. That matches what I've seen.

I still bring numbers, and every project starts with the current baseline. I can tell a leader what the metric is today, what I think we can do to it, and how we'll both know if I turn out to be wrong. I've watched leaders visibly relax when a proposal spells out the conditions where I lose.

Most of my influence work happens before there's anything to ask for, though. The thing that's gotten me the most mileage when it comes to influence is offering teams what I call AI 101: standing sessions where non-engineering employees bring their AI questions and use me as a sounding board for their own processes. How to use Claude and ChatGPT more efficiently, how to set up scheduled tasks, how to use artifacts, how to adjust memory. From my perspective it's just basic application education, but for them it's notably improving their work day, and it buys me a lot of credibility. Sachin Rekhi called influence without authority the most underrated PM skill, and his version starts in the same place mine does, with understanding what the other person is actually trying to get done. By the time I pitch a team a pilot, we already know each other.

I also keep the asks deliberately small, which is how I run projects anyway. A two-week pilot that needs part of one person's attention is an easy yes for a leader, where a quarter of their roadmap would need trust I haven't earned yet. And when a pilot works, the team that ran it gets the credit. It happened in their workflow and improved their number, so their people present the results. Those are the leaders who come back with ideas of their own, and after a year of this, most of my pipeline shows up that way.