Napster's Chief Technology and Product Officer, Edo Segal, recently shared a look into AI's "Orange Pill" moment, when Claude Code recalibrated this technology from a tool to, effectively, a limitless, tireless coding partner.

This series explores the practical applications of AI as a thought partner that does its best work when used to enhance, not replace, the human touch. It will also take a look at how to navigate shifting roles and responsibilities, avoid burnout, and acclimate to the new frictions of an AI-enabled workplace.

This installment takes that a step further. We've talked about using AI to clear the runway. Now let's work out what to do once you're on it.

Where friction falls

No matter how capable the model, AI cannot remove friction. Not entirely. It might pave the way a bit, but if you’re using AI properly, you’ll still hit some obstacles a ways down the road.

AI can handle the low-level friction of work — the syntax errors, the formatting passes, the boilerplate that fills hours without challenging you. What remains is hard in a different way: system design, conceptual integrity, the question of whether you're building the right thing at all. Those problems don't get easier just because the administrative layer got lighter. Those problems, as they always have, require the attention of an actual person.

And that’s a good thing. That means you’re treating the AI as a collaborator or a thought partner rather than a tool. It also means you’re putting your time and effort where it is most valuable.

There's a version of adapting to AI that focuses on efficiency: get the work done, and move on. The more interesting approach uses those same AI tools to not just check boxes but level up human ingenuity, provide a third-party perspective on big ideas, and achieve more.

AI as a thinking partner

Most people use AI for production: write this, summarize that, generate a few options. That's useful, but it's a narrow slice of what these tools can offer your thinking.

Research published in the Harvard Business Review found that AI tools, when used well, can meaningfully expand the range and originality of ideas a person generates. AI offers unmatched breadth: It generates many variations on a theme and draws connections across fields that a human mind, working alone, might not reach. The human contribution remains evaluation, judgment, and the kind of divergent thinking that produces genuinely novel ideas rather than well-organized familiar ones.

A Bain & Company survey of Fast Company 50 Most Innovative Companies found that AI is already reshaping early-stage concept development at organizations on the frontier of tech. That said, human-generated ideas remain significantly stronger in novelty, particularly for breakthrough thinking. You’re still the artist, but AI systems can broaden your palette in ways you might not expect.

Understanding that distinction changes the way a person approaches an AI companion and turns a tool for tasks that are already underway into a sounding board for potential game-changing work.

How to prompt your way into bigger thinking

The practical gap most people hit is knowing how to use AI for higher-level thinking rather than just more output. Napster Companions are built for this kind of collaboration: They have persistent memory, work conversationally, and are capable of building on context across sessions rather than starting cold every time. They can be, quite literally, a sounding board, one that can take a “yes, and” approach, be the blunt-but-necessary cynic who pokes holes in your ideas, and everything in between.

Here are five ways to use them for the hard thinking:

1. Start with the problem, not the deliverable. Instead of asking AI to draft a strategy deck, describe the underlying challenge you're trying to solve. "We're losing customers after the first month and I don't know why" will generate more useful thinking than "write a retention strategy." Let the framing develop in conversation before the output does.
2. Ask for the argument against your idea. One of the most underused prompts is simply: "What are the strongest objections to this?" or "What am I most likely getting wrong?" AI is good at steelmanning the other side, and pressure-testing your thinking before it reaches a real audience is one of the most practical things you can do with these tools.
3. Borrow from adjacent fields. Ask your AI companion how a similar problem has been solved in a completely different industry. A logistics challenge might have an elegant parallel in healthcare operations. A marketing positioning problem might find clarity through the lens of how a political campaign thinks about messaging. Cross-domain connection is something AI does naturally and humans tend to do only when they have time to wander.
4. Work through a block mid-project. The first part of this series covered AI's ability to solve the blank page problem. When you're deep into a project and the thinking has stalled, a conversation with an AI companion can surface assumptions you didn't know you were making or reframe a question in a way that moves things forward. Sometimes it just helps to articulate the block out loud to something that will ask a useful follow-up.
5. Build a feedback loop over time. Because Napster Companions retain memory across conversations, they can track how your thinking on a project has evolved and flag when you're circling back to ideas you've already rejected. That continuity is hard to replicate in a single session and genuinely useful for long-term work that develops over weeks or months.

The goal is to start by using AI to protect the conditions where your best thinking happens, then bring the platform into the loop to help elevate those ideas. Move the low-effort friction, center your workflow around the big picture, then come to the table again with better preparation, more room to explore, and a hard problem or two to solve.

You still have to decide if the solution is the right one. But that choice is easier with the kind of broadening of perspective that AI provides.

‍