---
title: "What is context decay? Definition | In Parallel"
description: "Context decay is the process by which recorded organisational knowledge becomes outdated while still looking current. Decisions are reversed, owners change and plans move on, but the documents, wikis and AI memories describing them stay as they were. Stale context then misleads people and AI tools more than having no record at all."
url: https://www.in-parallel.com/glossary/context-decay
---

[Glossary](https://www.in-parallel.com/glossary)

# What is context decay?

Definition

Context decay is the process by which recorded organisational knowledge becomes outdated while still looking current. Decisions are reversed, owners change and plans move on, but the documents, wikis and AI memories describing them stay as they were. Stale context then misleads people and AI tools more than having no record at all.

Also called: context drift, knowledge decay, stale context

Updated 22 October 2026

Every record starts true. The trouble is that most records have no way of knowing when they stop being so. A wiki page, a project brief or a briefing pasted into an AI assistant all carry the same authority on day ninety as on day one.

## How it happens

Context decays through change that does not reach the record. A decision is revisited in a meeting and reversed; the decision log is not updated. A project changes owner; the old name stays on the page. A client renegotiates; the account summary still shows the original terms. None of this is negligence. Updating records is extra work that rarely has an owner, and the people closest to the change already know.

AI tools speed up the damage. A model reading a stale page repeats it fluently, and an assistant’s saved memory can preserve an old fact long after everyone else has moved on.

## Why it matters

Decay creates work. People who cannot trust the record ask a colleague instead, and colleagues spend time re-explaining. The Coordination Tax Index 2026 (In Parallel, n=247 managers, Pollfish, June 2026) put re-explaining context at 4.6 hours a week per manager and searching for past decisions at 3.8. Decay also hides [execution drift](https://www.in-parallel.com/glossary/execution-drift), because an outdated plan offers nothing reliable to compare the work against.

## Context decay vs data quality

Data quality usually concerns whether a record was correct when entered. Context decay concerns records that were correct and became wrong with time. Validation at entry does not prevent it; only capture that keeps pace with change, and a visible age on every record, does.

## How In Parallel approaches it

In Parallel captures context where decisions happen, in meetings, so changes reach the record without a separate update step. Each decision carries when it occurred and the meeting it came from. [Pages](https://www.in-parallel.com/pages) are updated after the meetings that touch them and show the date they were last confirmed. Anyone who knows better can correct a record, including from an AI tool over [MCP](https://www.in-parallel.com/mcp). [How it stays current](https://www.in-parallel.com/how-it-stays-current) sets out each mechanism.

FAQ

## Common questions

**Why is stale context worse than none?**

With no record, people know to ask. With a stale one, they and their AI tools act on it with confidence. An assistant that cites last quarter's plan as fact produces an answer that looks right and is wrong, and nothing in the answer signals the problem.

**How can you tell whether context has decayed?**

Look for a date and a source. Context that shows when it was last confirmed and where it came from can be checked; context without either cannot. A record that is never corrected is also a warning sign, because real decisions change.

Related

[How it stays current](https://www.in-parallel.com/how-it-stays-current) | [Living plan](https://www.in-parallel.com/glossary/living-plan) | [Execution drift](https://www.in-parallel.com/glossary/execution-drift)

## Start with your next meeting.

Bring your team in. Use every surface, including every other team's. From €29 per user per month, billed monthly, no commitment.

[Get started](https://www.in-parallel.ai/)

[Book a demo](https://www.in-parallel.com/demo)

30-minute demo. We'll show you how meeting signals become a shared record of decisions, owners and risks.