Product
Conversation Continuity: The Missing Layer in Modern Workflows
We have systems for meetings, documents, and tasks. The state between them is usually carried by human memory.
A modern work stack remembers a lot.
Your calendar remembers where you were supposed to be. Meeting software remembers what was said. Documents remember decisions. Task managers remember assigned work. CRMs remember objects someone deliberately entered.
Then there is the conversation itself.
Someone asked a question. Someone promised an update. Ownership shifted. A stakeholder joined. A thread went quiet after a proposal. Nothing became a task, but the relationship still expects movement.
The continuity gap
Conversation continuity is the ability to preserve that unfinished state across time.
It is not a transcript. It is not a contact database. It is not “memory” in the sense of storing everything anyone ever said.
It is closer to remembering the unresolved edge of the relationship: who is waiting, what changed, what was promised, and whether the underlying work actually moved.
Most systems remember artifacts. Continuity remembers what is still unfinished between people.
Why tasks miss it
Tasks work well when someone explicitly names the work. Conversations often fail before that happens.
You read an email and think, “I’ll answer after this meeting.” A client says “let’s reconnect next quarter.” An investor asks one follow-up question. An intro lands while you are travelling. The next move is clear enough to a human in the moment and too informal to enter a system.
Then the moment passes.
That is not a productivity failure in the usual sense. It is a continuity failure.
Why agents make this more important
AI can make each individual message easier to process. It can summarize, draft, classify, and route.
But if every interaction is treated as a fresh object, the system still has the same weakness the human had: it can understand the current message without understanding what remained unfinished from before.
An agent that writes beautifully but forgets who owes the next move is still stateless in the way that matters.
Where Sidekick Log fits
Sidekick Log is an experiment in preserving this layer inside Gmail. Not every conversation deserves a task. Not every relationship belongs in a CRM. But some unfinished states are consequential enough that letting them disappear under newer email is expensive.
The product question is whether software can maintain that continuity quietly—and surface only the moments that deserve attention. That is the narrower idea behind email relationship tracking for Gmail.