---
{
  "title": "Agile delivery",
  "summary": "Use short feedback cycles to adapt toward a useful outcome.",
  "topic": "delivery",
  "level": "Beginner",
  "updated": "2026-10-12",
  "units": [
    2
  ],
  "aliases": [],
  "sources": [
    {
      "title": "The official Scrum Guide",
      "url": "https://scrumguides.org/scrum-guide.html"
    }
  ],
  "published": "2026-10-02",
  "editorialVersion": 2
}
---

## The idea

Agile approaches emphasize iterative delivery, collaboration, and responding to feedback. They do not remove planning, documentation, or accountability. Work is prioritized and refined as the team learns, while quality and security remain part of the delivery. Small increments can reduce uncertainty if they are genuinely usable and evaluated. Specific frameworks such as Scrum define additional roles and practices.

## In practice

A team improves account-security onboarding in small increments, tests user recovery needs, and adjusts the next work based on evidence.

## A common misunderstanding

Agile does not mean unplanned work or avoiding every document. Necessary controls should fit the workflow.

## Check your understanding

What makes an iteration valuable?

**Think it through:** It produces useful learning or an accepted improvement, not merely activity completed within a short period.


## Adapt the route while keeping a purpose

Agile approaches use feedback to improve both the product and the way people work. In a security project, that can mean releasing a small useful capability, observing its behavior, and revising the next step. It does not mean that requirements, planning, documentation, or accountability disappear. The amount and form of each should support the work rather than become an end in itself.

A useful increment needs a coherent purpose. Enabling a few log sources with an actionable alert and an owner can teach more than configuring many connectors that nobody investigates. Small delivery steps reduce the amount of uncertain work committed before feedback arrives. They still need appropriate quality and risk controls.

## A fictional alerting improvement

A team wants better visibility into administrator access. Instead of building a large dashboard first, it delivers one reviewed alert for a defined identity system and rehearses the response. Operators discover that the alert lacks the account's service owner, making escalation slow. The next increment improves that context before adding more alert types.

This feedback changes priorities without abandoning the objective. The team should explain the reason for the change and retain evidence of what it learned. Frequent changes caused only by competing requests are not automatically effective adaptation; they can prevent any useful result from reaching operation.

## Choose a learning step

Take a fictional plan to secure all staff laptops. Identify the smallest group and capability that would produce meaningful feedback while protecting participants. State what you want to learn, what quality conditions must hold, and which decision the result will inform. Avoid defining success only as completing tasks. The experiment should reduce a real uncertainty about the next delivery step.

## Keep exploring

- [[scrum]]
- [[kanban]]
- [[acceptance-criteria]]
- [[project-monitoring]]