By Alexandra Samuel

I've long been intrigued by what Doris Kearns Goodwin famously described as Abraham Lincoln's "team of rivals" strategy, where you get different people to advocate for their perspectives, to reach a better decision. It's a model that has inspired many leaders, because it helps get beyond the yes-men dilemma where nobody wants to challenge the boss.

But assembling a team of rivals takes effort: For one thing, you have to get everyone in a room -- or face an aggravatingly large video call. There isn't time to game out every decision in a room of 10 or 12 people. And if you're working in a small business or on your own, like I do, you may not have enough people to field a team at all.

At least, that was the situation.... until artificial intelligence came along. Now you can instantly convene a virtual, eclectic team of rivals to give you a range of approaches to any problem -- whether that's making your next career move, creating a new financial plan for your business, or handling a problem colleague. Pick as many people as you want, and make them smart, experienced and fractious; they don't have to worry about offending you or each other, and you don't have to worry about losing a good friend or colleague.

Unlike 10 humans competing for airtime in a stuffy conference room (and getting crankier as the meeting wears on), an AI can ask all those voices to vehemently disagree, and it can tell each expert to wait its turn to tell the others why they are horrifically wrong. That means you can watch the dispute unfold in an orderly way, and then ask the AI to neatly distill the conversation down to three or more options, with arguments for each option from the debate.

The disagreements also help address the problem of AI "sycophancy" -- AI's documented tendency to say whatever it thinks will make you happy.

Overcoming self-doubt

I've been using this team-of-rivals approach, and it has been helpful, both professionally and personally.

For instance, when I'm tackling a new topic to write about, I sometimes get overwhelmed by self-doubt: What if I get this story wrong? What if everyone can see I don't know what I'm talking about? What if I make an obvious mistake, or write something offensive?

These self-critical voices used to live entirely in my head, screaming at each other so loudly that I couldn't hear my own. Then I decided to ask AI to channel a half-dozen experts, each of which represents one kind of person I'm afraid will tear me to pieces once I'm in print -- including an organizational psychologist and a tech skeptic. I worked up the courage to pitch these virtual experts on a story idea I've been nervous about taking to a human editor.

What I heard was exactly what I was afraid of: "This is exactly the problem with tech journalists thinking they can waltz into specialized domains," one of my pseudo-academics said. "Where's her expertise?" demanded another.

My experts disagreed about what made me unqualified to cover the topic, and as they fought it out, I watched them defang one another's arguments -- which meant defanging my demons. "Her outsider perspective might be exactly what's needed to see patterns insiders miss," one AI voice argued. Hey, good point!

Watching these imaginary scholars debate my ideas and credentials was strangely liberating. Everything they said about me is something that I've already heard, and survived. Once I saw all the vicious critiques on my screen, instead of just in my own head, I remembered that criticism is simply part of publishing, and not a reason to avoid writing. The gremlins inside my head quieted down, and I had the mental space to work on my story.

What do I do next?

I also found my advisers helpful after I recently wrapped up a yearlong project, and found myself struggling with a case of the post-project blues. Should I jump back into another big project? Should I take time off? How could I get my mojo back?

I asked AI to spin up a team of advisers who could duke it out. It even gave them names, including Dr. Marcus "Metrics" Johnson, a 38-year-old productivity consultant who is skeptical of feelings-based approaches; Brad "Always Be Closing" Thompson, a 45-year-old business consultant who thinks rest is weakness; and Prof. Simone "Why Mojo" Beauvoir, an existential philosopher who questions everything.

The debate quickly polarized between the hard-driving achievers and the touchy-feely regenerators. "While you're navel-gazing about 'mojo,' someone else is landing your next client," said one voice. But another voice pushed back: "This is exactly the BS that burns people out! Not everyone's brain works on your capitalism schedule!"

Seeing these extremes made it clear I wasn't prepared to follow either direction: No powering through, and no vacation for me, either. And definitely no "daily 15-minute 'grief sits' to process the loss of project purpose," as one AI faction suggested. Who would have predicted I'd be the one to call an AI too sentimental?

My reaction to all these proposals helped me see that I needed an in-between solution, and I found a bit of inspiration in the portion of the AI debate that suggested organizing daily dopamine hits -- more dinner dates with friends, quirky craft projects and an infusion of live theater.

Conflict with colleagues

To see if my team of rivals could have helped with a problem from the past, I asked about something I deal with from time to time as a consultant: landing in the middle of a conflict among colleagues. Each person on the team typically has a different vision for the project we're working on, and I am stuck trying to keep everyone happy.

Watching the AIs (including a master of corporate politics and a consultant with experience in similar conflicts) argue over the right solution made me realize that it isn't some personal failing when I get caught up in internal politics: Workplace conflicts are complicated, and the fact that a team of 10 AIs couldn't come up with even one realistic solution made me feel better about how I typically muddle through. But the exercise wasn't a total bust, because their debate got me thinking about options that hadn't occurred to me -- like asking for nonmonetary compensation for my extra hours, or documenting the extra work as a one-time accommodation.

Consulting an AI team of rivals doesn't mean I'm always making the right decision. When I turn to human colleagues for advice, they'll inevitably offer ideas that the AIs have failed to note. But getting feedback from a human is even more useful if I've first road-tested my ideas with a panel of AIs, because I've already anticipated and addressed the most obvious objections. That means I can ask my human colleagues to do what humans still do best: Point out the glitches and opportunities that are so far outside the box, even the biggest data set on earth can't predict them.

Alexandra Samuel is a technology researcher and co-author of "Remote, Inc.: How to Thrive at Work...Wherever You Are." She can be reached at reports@wsj.com.

(END) Dow Jones Newswires

August 29, 2025 09:00 ET (13:00 GMT)

Copyright (c) 2025 Dow Jones & Company, Inc.