The Single-Tasking Framework with AI: How the One Thing Lock Works

Every focus framework tries to solve the same problem: the gap between intending to focus and actually doing it.

Most frameworks address this with time. Thirty minutes, twenty-five minutes, ninety minutes—the bet is that a bounded interval makes it easier to commit. Time bounds help. But they do not solve the underlying problem, which is not about duration. It is about open loops.

The One Thing Lock is a framework built around that insight.

The Problem Frameworks Usually Miss

When you sit down to work on one thing, you are rarely thinking about only one thing. You are thinking about the task in front of you and the three tasks waiting, the email you need to send, the decision you are supposed to make this afternoon, and the thing you said you would remember to do.

Sophie Leroy’s 2009 research on attention residue demonstrated that this is not a personality flaw. It is how working memory operates under conditions of incomplete tasks. Until an open loop is closed—either completed or offloaded into a trusted system—it continues to consume attentional resources.

Pomodoro addresses this with brevity (just wait 25 minutes). Deep work scheduling addresses it with long blocks and deliberate practice. Neither directly addresses what to do with the queue.

The One Thing Lock addresses it with offloading.

The Three Phases in Detail

Phase 1: The Lock (5 minutes)

The Lock is a structured pre-session protocol, not a casual warmup. Its purpose is cognitive preparation: closing open loops before you need to focus, not while you are trying to.

Step 1 — Define the task output, not the activity.

There is a meaningful difference between “work on the proposal” and “complete the scope section of the proposal (two pages).” The first is an activity. The second is an output with observable completion criteria.

Output-defined tasks do three things: they direct attention precisely, they make procrastination harder (there is no ambiguous middle ground to occupy), and they give you unambiguous closure when you are done.

Step 2 — Identify the pending queue.

Before you start, spend sixty seconds listing everything your brain is holding: incoming tasks, pending replies, things you are worried about forgetting. Do not filter or prioritize—just list.

This step acknowledges that your working memory is already loaded. The list externalizes the load.

Step 3 — Hand the queue to your AI.

This is the mechanism that makes the One Thing Lock structurally different from standard focus advice.

One Thing Lock setup:
Task this block: [output-defined task]
Queue for after: [your list]
Duration: 45 minutes
Return briefing: yes
Task-focus reminder: yes

Once the AI holds the queue, your brain receives a signal it can act on: the pending items are safe. They will not be lost. You do not need to monitor them. That signal is the difference between performing focus and actually having it.

Phase 2: The Lock-In (45 minutes)

The lock-in has two rules:

1. Work only on the stated task.
2. Use AI only for task-relevant support.

The second rule deserves emphasis. AI assistants are designed to be responsive and engaging. An open AI chat window during a focus session is an ambient distraction waiting to be activated. The discipline is to use it directionally: “help me with this specific part of this specific task” rather than “let me see what this brings up.”

Managing in-session intrusions:

New thoughts will arrive. That is not focus failure—it is how cognition works. The protocol for handling them:

• Write it on a notepad (physical is better than digital here—it is less likely to cascade into checking something).
• Continue working.
• Add it to the queue during the unlock.

Do not evaluate whether the intrusion is important enough to break focus for. That evaluation itself costs focus. Everything goes to the notepad; everything gets assessed after the block.

The 45-minute duration:

Forty-five minutes is a practical default, not a magic number. It sits within the ultradian rhythm window that Nathaniel Kleitman and Peretz Lavie described in sleep and waking rest-activity research—cycles of approximately 90 minutes with natural peaks and troughs. A 45-minute block captures the ascending alertness arc before the natural attention trough arrives.

For individuals who find 45 minutes too long initially, 25 minutes is a reasonable starting point. For those running deep analytical or creative work who find 45 minutes too short, 60 or 75 minutes can work—provided the queue has been properly offloaded and the task is output-defined.

Phase 3: The Unlock (5–10 minutes)

The unlock is the most underrated phase.

Without it, you exit the lock-in still holding the previous task mentally (is it done enough? should I go back to it?), plus the re-emerging queue, plus whatever new things arrived during the session. The cognitive transition cost—what David Meyer’s research identifies as executive reconfiguration—hits you unprepared.

The unlock structures that transition.

Block complete. Task status: [done / partially complete — here's what's left]
New captures from block: [notepad items]
Please give me the queue briefing and any new items I need to add.

Your AI returns the queue, you add your notepad captures, and you make one explicit decision: what is next. Not what feels most urgent in the moment. What you would choose if you were thinking clearly.

That decision point is the whole value of the unlock. It interrupts the reactive chain—the tendency to just respond to whatever is loudest after a focus session—and replaces it with an intentional choice.

Four Adaptations for Different Work Contexts

The One Thing Lock is a framework, not a rigid prescription. Here is how to adapt it.

Interrupt-driven roles (customer success, operations, support)

Reduce block length to 25 minutes. Define a narrow, P0 interruption criterion with your manager or team in advance. Everything below that threshold goes to the queue. Most interrupt-driven workers find that 80 percent of what feels like urgent interruption is actually tolerant of a 25-minute delay.

Long-form creative work (writing, design, research)

Extend blocks to 60 or 75 minutes. Use the Lock more carefully—spend the five-minute setup drafting a mini-outline or brief for the creative session, not just a task definition. This primes the creative context before you enter the block.

Collaborative knowledge work (product management, consulting)

Run locks only for the portions of your work that are individual—strategy drafts, analysis, document creation. Use the unlock not just for queue review but for scheduling any collaboration that emerged from the block.

High-meeting days

Run micro-locks in the gaps between meetings—25 to 30 minutes. The key discipline here is not expanding the gap into email/Slack time. Treat every gap as a potential lock-in, even a short one. The cumulative effect across a fragmented day is significant.

How AI Changes the Framework’s Leverage Point

Traditional focus frameworks treat the human as the sole load-bearing element. You must remember the queue, manage the distractions, track the time, and make the transition decisions. That is a lot to hold.

The One Thing Lock redistributes that load. AI holds the queue. A timer handles the duration. The framework handles the transition structure. You are responsible for the work itself.

This redistribution matters because cognitive load affects not just quantity of output but quality. When your working memory is managing queue anxiety simultaneously with the task at hand, the task gets a fraction of your capacity. When the queue is externalized, the task gets most of it.

Beyond Time (beyondtime.ai) integrates this pattern directly—its session-based planning interface is designed to capture the queue before a focus block and surface it cleanly at the end, without requiring you to construct the prompts from scratch each time.

The Mistake That Undermines the Framework

The single most common implementation failure is treating the Lock as optional.

“I know what I’m working on, I don’t need to write it down or tell the AI.” This misunderstands what the Lock does. It is not documentation. It is the act of cognitive commitment. Writing the task definition and sending it to the AI is a behavioral cue that initiates the mental mode of focused work—analogous to what Peter Gollwitzer’s research on implementation intentions calls the “if-then” planning trigger.

The five minutes feel like overhead. They are not. They are the mechanism.

Run the full three-phase protocol on your next focused task—including the five-minute Lock even if it feels unnecessary—and compare the quality of focus to your last underprepared session.

Related:

• The Complete Guide to Single-Tasking with AI Support
• 5 Single-Tasking Approaches Compared
• The Science of Single-Tasking
• How to Single-Task with AI Support

Tags: single-tasking, One Thing Lock, focus framework, AI productivity, attention residue