One of the more fascinating conversations I’ve had in a long time didn’t start with AI, agents, or customer experience buzzwords. It started with a LinkedIn post about preserving intellectual knowledge.

That article, “The Vanishing Wisdom: The Knowledge Tapped Between Two Ears,” was written by Annemarie Pucher, CEO of ISIS Papyrus Group, and it immediately got my attention. Not because it was provocative, but because it articulated something I’ve watched enterprises struggle with for years, and have experienced firsthand when experts with decades of knowledge and judgment retire or walk out the door.

What followed was a wide-ranging conversation that made one thing very clear. Papyrus Software hasn’t been chasing where the CCM market says it’s going. They’ve been quietly building for where enterprises eventually have to go.

Before “Agents” Had a Name

Papyrus came out of the same origins as many long-standing CCM vendors in the late 1980s: mainframes, printers, Xerox 9700s, and direct-to-output architectures. But instead of staying anchored to output, they moved early into something most of the market resisted at the time: process.

A document without a process, as Annemarie put it bluntly, is worth nothing.

By the early 2000s, Papyrus was already closing the loop between inbound and outbound communications, integrating data capture, classification, and automation while most organizations insisted those worlds should stay separate.

Then came a detail that feels almost surreal in hindsight. In 2008, Papyrus patented what they called a user-trained agent. At the time, nobody really knew what to do with agents. The industry didn’t have language for it. There was no category to put it in.

Today, everyone is talking about agents.

The Enterprise as a Living System

I really love the way Annemarie describes the enterprise. Not as a stack or a collection of tools, but as a living system. More like a doctor diagnosing a patient than an architect assembling components.

Documents, data, and processes aren’t separate domains in this view. They are indicators. Evidence of decisions. Traces of how work actually gets done, including all the exceptions, workarounds, and human judgment that never make it into formal documentation.

TreelinePress is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

This perspective matters because it reframes the problem enterprises are trying to solve today as they rush to feed LLMs and AI systems. The issue isn’t just automation. It’s visibility. It’s understanding why decisions are made, not just what happened.

That’s also where knowledge loss becomes painfully real. When experienced people leave, the rules don’t just disappear from systems. They disappear from context. From memory. From the space between two ears.

Why This Matters Right Now

It’s impossible to have this conversation in 2026 without talking about AI. But what made this discussion different with Annemarie was how grounded it was.

AI doesn’t fail because models aren’t powerful enough. It fails because enterprises lack context. Without shared language, visible rules, and clear process intent, intelligence becomes guesswork.

Papyrus’ focus on adaptive processes, business language, and ontology isn’t academic. It’s foundational. AI needs something to reason over. It needs enterprises that can describe themselves clearly and consistently.

In that sense, Papyrus wasn’t early to AI. They were early to what AI actually needs in order to work effectively.

Andy’s Take

You often hear about AI as if it is something you can bolt on at the end of an outdated process. A magic bullet that will make all things possible.

This conversation made the opposite case. Intelligence starts with understanding how work actually gets done. It starts with capturing judgment, exceptions, and intent while that knowledge still exists. AI doesn’t replace that work. It depends on it.

Too many transformation efforts fail because enterprises try to automate what they don’t truly understand. They digitize outputs without addressing the processes and decisions underneath. Then they wonder why AI projects struggle to deliver meaningful results.

Papyrus’ story is a reminder that real progress often looks boring before it looks brilliant. It doesn’t announce itself with buzzwords. It shows up quietly, years ahead of demand, waiting for the rest of the market to catch up.

Get full access to TreelinePress at treelinepress.substack.com/subscribe