For decades, business leaders faced an impossible choice when it came to understanding their markets, customers, or prospects: invest heavily in deep research or settle for surface-level insights you could afford.
Want to truly understand a customer's business before a critical meeting? That meant hiring a research team, spending days gathering intelligence, and coordinating briefings. Need qualitative depth to understand customer motivations? Budget for weeks of interviews, transcription, and analysis. Looking for quantitative scale to validate patterns? Pay for large sample surveys and wait for results.
The constraint was real: depth required resources most teams didn't have.
This connects to two of Amazon's Leadership Principles that rarely get discussed together, but in the AI era, they've become inseparable: Dive Deep and Frugality.
Dive Deep states: "Leaders operate at all levels, stay connected to the details, audit frequently, and are skeptical when metrics and anecdotes differ. No task is beneath them."
Frugality states: "Accomplish more with less. Constraints breed resourcefulness, self-sufficiency, and invention. There are no extra points for growing headcount, budget size, or fixed expense."
At first glance, these principles seem contradictory. How do you operate at all levels and stay connected to details without spending significantly? How do you audit frequently and remain skeptical of surface-level metrics while being frugal with resources?
For most of my career, this was the fundamental tension in research and customer understanding. You chose your constraint: either time, money, or depth.
AI doesn't just make research faster or cheaper - it eliminates the trade-off entirely.|Consider what "qualitative versus quantitative" meant until very recently. Qualitative research gave you depth: rich interviews, nuanced understanding, genuine customer voice. But you needed to use humans to lead it, and could afford maybe 8-12 interviews. Quantitative research gave you scale: hundreds or thousands of responses, statistical significance. But you lost the richness, the "why" behind the numbers.
Business leaders had to choose. Do we invest in understanding a few customers deeply, or do we gather surface-level data from many? The decision often came down to budget, timeline, and what you could justify to stakeholders.
The LSE researchers Friedrich Geiecke and Xavier Jaravel demonstrated what's now possible: AI-led interviews that generate 142% more detailed responses than traditional text fields, can be conducted with thousands of participants simultaneously, and produce insights rated comparable to human expert interviews. What used to require choosing between depth OR scale now delivers depth AND scale.
This isn't just faster research. It's fundamentally different research. You can now dive deep with the rigor of operating at all levels - while being radically more frugal with time, budget, and headcount.
Smaller teams can now do work that previously required large research departments. The constraint that forced the trade-off - human capacity - no longer binds in the same way. Every person on your team can now operate with dozens of AI agents helping them dive deeper into details that would have been impossible to access before.
Think about what a single person equipped with the right AI tools can accomplish today:
Deep research capabilities through ChatGPT's research mode, Gemini's deep analysis, or Perplexity's source synthesis
Qualitative analysis at scale through platforms that conduct and analyze hundreds of interviews
Synthesis across multiple data sources - documents, conversations, market intelligence - that would have taken a team weeks to compile
The multiplication factor is stunning. What required a six-person research team working for weeks can now be done by one person with the right AI tools in days - often hours.
Last week, I talked with a sales division at a B2B services company. They regualrly have meetings with C-level executives at major potential clients. Previously, this preparation would have meant: briefing from the account team, generic company research from an analyst, and the salesperson's intuition about how to position.
Instead, I showed them how to chain three custom GPTs:
Deep research agent: Gathered comprehensive intelligence on the executive, their company, recent strategic initiatives, and industry challenges
Messaging architect: Crafted role-specific, personalized messaging that connected their solution to the executive's actual priorities
Scenario simulator: Role-played the meeting, anticipating objections and questions based on the executive's background and communication style
The entire process took 45 minutes - including the role play and feedback. Two years ago, this level of preparation would have required multiple people working for days - and most companies simply wouldn't have done it because the ROI didn't justify the resource investment.
That's what being frugally deep looks like in practice: accomplishing sophisticated, detailed work with minimal resources because AI amplifies human capability rather than replacing it.
For leadership - this isn't about tools. It's about rejecting false choices. For years, leaders accepted certain trade-offs as inevitable: "We can be thorough OR we can be fast." "We can have depth OR we can have scale." "We can be rigorous OR we can be frugal."
The best leaders question whether constraints are real or just conventional wisdom. They recognize when technology fundamentally changes what's possible and adjust accordingly.
AI doesn't just make the old trade-offs more efficient - it makes them obsolete. You don't have to choose between diving deep and being frugal anymore. You can be both, simultaneously, because the constraint that forced the choice (human capacity and time) has been amplified exponentially.
This requires a different approach to team structure, tool adoption, and how you evaluate what's "good enough" in research and customer understanding. If a single person with the right AI tools can accomplish what previously required a team, what does your organization do with that multiplied capacity?
Look at the trade-offs your team currently accepts as inevitable. Which ones are still real constraints, and which are just habits from a world where human capacity was the limiting factor? Pick one area where you've been settling for "good enough" research or customer understanding because deeper analysis seemed too expensive or time-consuming. Test whether AI tools can eliminate that trade-off entirely.
The leaders who thrive in the AI era won't be those who do the old work faster - they'll be those who recognize which old constraints no longer exist.