Ask anyone who has returned from a long leave and they will describe the same experience. You come back, and there is a stack of meeting notes waiting. Maybe a project document or two. Maybe a Confluence page someone remembered to update.

And none of it tells you what you actually need to know.

It tells you what was decided. It does not tell you why that option won over the other ones. It does not tell you what concerns got raised and then quietly set aside. It does not tell you what the team almost did before they changed direction, or what finally brought a hesitant stakeholder around.

That context lives in the people who were in the room. When those people are not available, it is just gone.

I have watched this happen on enough projects to know: the documentation gap is rarely about effort. Teams document plenty. The problem is that documentation captures outputs, not thinking.

What I was actually dealing with

Earlier this year, a MarTech leader I work closely with went on maternity leave. Right in the middle of a complex, cross-functional implementation. We're talking a full Iterable to Braze migration. Big project. Multiple workstreams. Decisions being made every week across technical teams, business stakeholders, and vendors. Her team was responsible for leading it.

While she was out, I was filling in for her. All of it. Day-to-day MarTech project management. Leading the migration. Managing her team.

There was a lot happening. New decisions, new context, new wrinkles in the project every week. And I kept thinking: by the time she comes back, she's going to need to catch up on months of things that don't fit neatly into a status update. Not just what happened, but why. Which calls changed direction and what drove that. What the team tried that didn't work. Who pushed back on what and what we did about it.

You can't put that in a handoff doc. And the traditional answer, document everything before someone leaves, wasn't even an option here. She was already gone. New things were happening every day.

I didn't need a document. I needed something that could learn as the project moved.

The goal was not a handoff document. It was a companion she could come back to and ask anything of, in plain language, and get an answer grounded in what actually happened while she was away.

What I was trying to build

What I built while she was out

I built an AI project agent specifically designed to capture, organize, and preserve the context that was accumulating in real time. Every week, as things happened, the system was updated. It was not built before she left. It was built during her leave, as a living record of everything she would need to know when she returned.

How the system worked
🎙️
Record every key meeting Not all meetings, the ones where decisions were made, tradeoffs were weighed, or direction shifted. The recording is what makes everything else possible.
📝
Generate transcripts automatically No manual note-taking required. The transcript captures everything that was said, including the side conversations and the questions that got asked before the final answer.
🗂️
Create structured summaries of decisions and rationale Not just what was decided, who raised it, what alternatives were on the table, what concerns got voiced, and what ultimately drove the call. The why, not just the what.
🧠
Organize everything into a queryable AI workspace The summaries, the transcripts, the decision logs, all of it goes into one place where anyone on the team can ask questions and get answers grounded in actual project history.

That last part is where things changed. Not the transcription, that's been possible for a while. The difference was what happened once all that captured knowledge became something you could talk to.

Instead of searching through documents hoping you'll find the right one, anyone on the team could just ask:

Questions the system could answer
Why did we decide to go with this approach instead of the alternative?
What concerns did the stakeholder group raise in the March review?
Who owns the data migration workstream, and what was the last decision made there?
What changed between the original plan and where we landed?
What follow-up actions came out of the vendor call last month?
What were we worried about before we made this call?

The AI could answer all of these, not from a summary someone wrote after the fact, but from the actual conversations where those things were said.

What it actually did

When decisions came up that echoed something the team had already worked through, the reasoning from weeks earlier could be surfaced instead of relitigated. Stakeholders stopped getting asked the same questions repeatedly because the answers were already captured somewhere findable.

And when the leader came back from leave, she was not spending her first two weeks in status meetings trying to reconstruct what had happened. She could ask the system. She got answers. She was back in the work within days instead of weeks, because the context she needed was already organized and waiting for her.

The shift that mattered

The system did not just help someone return from leave. It changed the cost of not being in the room. When institutional knowledge lives in a system instead of in people's heads, absence stops being a gap. Everyone on the team has access to the same history, regardless of when they joined, whether they were in that meeting, or how long they have been gone.

Before the knowledge system
  • New hire required multiple catch-up calls to understand project history
  • Decision rationale lived in meeting memories, not anywhere findable
  • Returning leader faced weeks of status meetings to reconstruct context
  • Teams relitigated decisions because the reasoning was not documented
  • Stakeholders were asked the same questions over and over
  • Knowledge accumulation during the leave had no home
After the knowledge system
  • Returning employee was productive within days, not weeks
  • Decision history, including the why, queryable by anyone
  • New team members self-onboarded through the knowledge base
  • Prior reasoning surfaced automatically when similar questions arose
  • Stakeholders freed from repeating context they'd already shared
  • Continuity built into how the project operated, not bolted on at the end

Where this matters beyond one project

This started as a maternity leave solution. It didn't stay one.

The same system, record, transcribe, summarize, organize, query, applies anywhere knowledge loss is a risk. Which turns out to be almost everywhere.

🔄
Employee transitions When someone changes roles or leaves, their institutional knowledge stays behind instead of walking out with them.
🚀
New hire onboarding Instead of weeks of catch-up calls, new people can ask questions and get answers grounded in actual project history.
🤝
Consulting engagements Context and rationale built up over months don't evaporate when a consultant's engagement ends or the team rolls over.
📈
Long-running transformations Multi-year programs accumulate enormous institutional knowledge. This makes it accessible throughout, not just for the people who were there from the start.
🏗️
Reorgs and restructures When teams change, the history of why things were built the way they were stays available to the new structure.
🌍
Cross-functional initiatives When no single person has full visibility, a shared knowledge base gives everyone access to the same project truth.

I have spent more than 20 years working on enterprise marketing technology. The pattern I see over and over is that organizations invest heavily in their systems and their processes, and almost nothing in preserving the thinking behind them. That thinking is what makes the systems make sense. Without it, every transition (every new hire, every leadership change, every reorg) starts from a little less than it should.

AI gives us a way to fix that. Not by replacing the conversations, but by making sure they do not disappear.

The most valuable thing AI can do in enterprise work is not produce faster outputs. It is carry the context forward so that the next person in the room does not have to start from scratch.

Thinking about knowledge continuity?

Let's talk through what this could look like for your team

If you're managing a complex project, a long-running transformation, or a team that's about to go through a transition of any kind, I'd love to think through this with you. It's one of those solutions that sounds simple once you see it, and makes you wonder why you weren't doing it all along.

Let's connect See more of my work