What 100+ user interviews taught us about how knowledge workers actually think

What we learned about the nuanced reality of deep knowledge work

Over the past few months, we've conducted over 100 user interviews with professionals across industries: think tank researchers developing policy briefs, VCs writing investment memos, strategy teams crafting recommendations, graduate students working on dissertations, and executives synthesizing market intelligence.

We started these conversations to understand how to build better tools for knowledge work. What we discovered was that the reality of how these professionals think and create insights is far more nuanced than commonly understood.

These conversations fundamentally shaped how we're approaching knowledge synthesis at Liminary: not as another storage solution, but as a system that works with how thinking actually happens.

Here's what they really told us about their craft, insights that reveal the sophisticated nature of thinking work and why one-size-fits-all solutions often miss the mark.

1. The friction paradox: why smart people are thoughtful about AI adoption

Our first insight: the most experienced knowledge workers often approach AI tools carefully, not because they're resistant to new technology, but because they understand something fundamental about how deep thinking works.

Friction is a feature, not a bug. The researchers, analysts, and strategists we spoke with need to actively engage with content to generate genuine insights. The process of note-taking helps internalize knowledge, and writing out key points helps convert information into understanding.

This isn't stubbornness, it's wisdom. These professionals understand the difference between AI-assisted efficiency and the deep cognitive work that produces breakthrough insights.

The 80% challenge: Multiple users described a phenomenon where AI gets them "80% of the way there," which sounds helpful until you realize that the last 20% often requires completely rethinking the approach. They end up spending time editing AI-generated content rather than doing their own thinking.

2. Writing is thinking (and editing AI isn't)

"For me, writing is thinking." - Think tank researcher

This phrase appeared in interview after interview, revealing something important about how knowledge work actually happens. These professionals don't write to communicate thoughts they already have, they write to discover what they think.

The blank page preference: Users consistently told us they'd rather face a blank page than edit someone else's work (including AI's). Starting from another’s reference point requires significant cognitive effort to adopt and modify. Many professionals would rather start from scratch than feel backed into someone else's conceptual corner.

Physical writing matters too: Even in our digital age, many of these professionals still rely on pen and paper for their most important thinking. The tactile experience and pace of handwriting activates different cognitive processes that typing can't replicate.

3. The social memory system (that search can't access)

"I know this person said it. I know it's in my meeting notes, I just don't know what date or which meeting." - Strategy consultant

Researchers, analysts, and strategists don't just remember information: they remember the social context around it. Who mentioned it in a meeting. Which expert shared it. What the author's perspective was. This social metadata is crucial for evaluating and connecting information, yet it's often invisible to traditional search approaches.

One VC described the challenge of trying to trace insights back to their sources, especially when recency bias makes it easier to remember more recent conversations than older, potentially more relevant ones.

4. The template paradox

Every knowledge worker we spoke with relies on templates and frameworks, yet they also know that their most valuable insights emerge outside these structures. Strategy consultants have frameworks for market analysis. VCs have templates for investment memos. Researchers have methodologies for literature reviews.

But here's the paradox: the content that breaks the mold tends to be most unique and helpful. As one consultant noted, each project is "bespoke" despite following general patterns. The innovation happens in the spaces between the template sections.

5. The personal organization paradox

One of our most surprising findings: humans in roles requiring deep thinking feel guilty about not having sophisticated personal knowledge management systems. The reality? Most don't have them, and those who do often question whether the maintenance effort is worthwhile.

There is a tool proliferation trap. These professionals hate having their information scattered across multiple tools, yet they continue adopting new ones. One researcher summed up the frustration: they're "bastardizing about five different products and tools because none of them quite do exactly what I need."

No single tool seems adequate for all information types or contexts, leading to an expanding toolbox despite the frustration it causes.

6. Audience management: the hidden layer of knowledge work

Researchers, consultants, and analysts aren't just focused on ideas, they're constantly thinking about how their audience will receive and act on information. Whether they're strategy consultants or think tank researchers, they're strategizing about how to trigger specific outcomes from their readers.

This audience awareness shapes every aspect of how they research, synthesize, and present information: a sophisticated layer that's often overlooked in tool design. One strategy consultant described rewriting the same content multiple times, worried about misinterpretation in text form.

7. The search problem that persists

Despite all our technological advances, finding information you've seen before remains one of the most consistent pain points. We spend 30% of our time just looking for information we already found.¹

One private equity professional described the familiar frustration: hunting through emails to find a company deck, checking if it was saved to the shared drive, then navigating through multiple systems just to locate a document they knew existed somewhere.

This isn't a simple search problem, but rather it's about context, connection, and the complex ways human memory actually works.

What this means for the future of knowledge work

These insights reveal something important: the most experienced professionals approach new tools thoughtfully because they understand how thinking actually works. They're not avoiding AI, they're looking for AI that works with their cognitive processes rather than against them.

The future of knowledge work tools isn't about replacing human thinking, it's about amplifying it while preserving the cognitive engagement that makes insights possible. The most experienced knowledge workers want AI that amplifies their thinking rather than replacing it. That's exactly the gap we're working to bridge.

The best researchers, analysts, and strategists want AI that handles the mundane parts of information management so they can focus on what humans do best: making unexpected connections, developing original arguments, and generating novel insights.

As one researcher perfectly summarized their ideal, "I like research, I don't want it to do that part. What I want is a tool that helps me think better, not thinks for me."

Sources

¹ CDP Institute, "Knowledge Workers Lose 30% of Time Looking for Data: Forrester Study," https://www.cdpinstitute.org/news/knowledge-workers-lose-30-of-time-looking-for-data-forrester-study/