David Allen published Getting Things Done in 2001, when the average inbox was a fraction of today's and "capture" meant writing things on paper before they escaped. Twenty-five years later the book's core insight has only gotten more true: your mind is for having ideas, not holding them. What's changed is where the ideas arrive. For most professionals, the answer is email, at a rate of around 120 a day.
That makes email the natural home of GTD in 2026, and also the place where the method most often breaks. Let's map the five steps onto a real inbox and be honest about where the friction is.
Step 1: Capture (your inbox already does this)
GTD's first step is collecting everything that has your attention into trusted inboxes. Email is capture running on autopilot: colleagues, clients, and tools deposit their requests directly. You get this step for free. The catch is that free capture means unlimited capture, which puts all the pressure on step two.
Step 2: Clarify (this is where the work is)
Clarifying means asking of each item: what is this, and is it actionable? In email terms, every message resolves into one of five buckets:
| Bucket | The email says | It becomes |
|---|---|---|
| Next action | "Can you review the draft?" | A verb-first task with a date |
| Waiting-for | You asked, they haven't answered | An entry on your waiting-for list |
| Calendar | "The review call is Thursday at 2" | A calendar entry, not a task |
| Reference | The signed contract, attached | Filed, searchable, out of sight |
| Trash | Everything else | Archive, unsubscribe |
Allen's two-minute rule applies during clarifying: if the action takes less than two minutes, do it on the spot, because tracking it would cost more than doing it. Short confirmations and quick answers get handled inside the processing session itself.
Step 3: Organize (two lists carry the load)
The next-actions list and the waiting-for list do most of GTD's work for email users. The waiting-for list deserves special attention because email generates waiting-fors constantly: every question you send, every task you delegate, every proposal that leaves your outbox. Untracked, these depend on memory, and memory chases late or never. Tracked, a Friday review catches everyone who's gone quiet.
Step 4: Reflect (the weekly review)
Once a week, sweep the inbox to zero, walk the next-actions list, and chase the stale waiting-fors. Thirty minutes, calendar-protected. Allen calls the weekly review the critical success factor of the whole method, and for email users it's also the safety net that catches whatever a chaotic Tuesday dropped. Our inbox zero guide covers the reset routine in more detail.
Step 5: Engage (do the work from the list, not the inbox)
Work from your clarified lists, never from the raw inbox. The inbox sorts by arrival and rewards whoever emailed most recently. Your list sorts by what you decided matters. Working from the inbox is how the loudest thread wins the day.
Where GTD breaks, and what fixes it
Be honest about the failure mode: clarifying is a per-item cost, and email volume grows without asking. At 30 emails a day, hand-clarifying is a pleasant ritual. At 120, it's 40+ minutes of sorting before any actual work begins, and one busy week leaves a backlog that makes the whole system feel like debt. Most abandoned GTD practices die exactly here, at step two, under volume.
This is the step AI now runs well, because clarifying email is reading and categorizing, which is pattern work. Pidgy reads your Gmail or Outlook inbox and performs the clarify-and-organize pass automatically: requests become verb-first next actions, sent mail awaiting replies becomes your waiting-for list, pure information gets filed as reference. Everything stays in one place inside Pidgy, with calendar planning on the roadmap. You review the sorted result and engage, which was always the part that needed you.
Purists sometimes ask whether automated capture-and-clarify still counts as GTD. Allen's own answer to tooling questions has always been that the method is about a trusted system, not a specific tool. If every open loop is captured, clarified, and reviewable, the system is trustworthy. Whether a human or a machine did the sorting was never the point.
