Single-Tasking with AI Support: Your Questions Answered

The Research

Q: Is the case against multitasking actually strong, or is it overstated?

The core findings are robust. Sophie Leroy’s 2009 attention residue research, Clifford Nass’s Stanford study on heavy multitaskers, and David Meyer’s task-switching cost research all point in the same direction: switching between tasks carries cognitive costs, and those costs are not reduced by practice—they may be worsened by habitual multitasking.

Where the evidence is less settled: the precise magnitude of the costs in real-world knowledge work (lab estimates may differ from field conditions), whether multitasking has different costs for different task types (the evidence is strongest for language-based and analytical tasks), and the causal direction in correlational studies.

The honest summary: the evidence is strong enough to take seriously and act on. It is not so precise that you can calculate exactly how much productivity you are losing.

Q: My manager says multitasking is just a reality in our industry. Is there any evidence they are wrong?

The evidence that multitasking impairs cognitive performance is not industry-specific. It applies to knowledge work broadly. What your manager may mean is that switching contexts is unavoidable—which is often true—not that switching is costless. Accepting that switching will happen is different from accepting that it should happen as often as possible.

The productive reframe is: given that some switching is unavoidable, how do you minimize unnecessary switching? That is a goal most managers would support.

Q: What is attention residue and why does it matter more than just switching speed?

Attention residue, as described by Sophie Leroy, is the cognitive processing of a previous task that continues after you have switched away from it. The residue is not about how long the switch takes—it is about where your attention goes after the switch.

If you switch from an incomplete analysis to a client call, part of your cognitive processing remains on the analysis. You are on the call in body but not fully in attention. The call gets a fraction of your capacity. The error rate on both activities increases.

This matters more than raw switching speed because it means the cost of a task switch extends well beyond the moment of the switch itself. A thirty-second task switch can produce fifteen minutes of degraded performance.

The Protocol

Q: What is the One Thing Lock, and how is it different from just focusing?

The One Thing Lock is a three-phase protocol: a structured five-minute setup (the Lock), a 45-minute focused work session (the Lock-In), and a five-to-ten-minute structured handoff (the Unlock).

The difference from “just focusing” is the open-loop management mechanism. Most focus advice tells you to resist distraction. The One Thing Lock removes the primary source of self-generated distraction—the mental queue—before the session starts by offloading it to an AI. You do not need to resist the urge to monitor the queue because you have already handled it.

Q: Why 45 minutes? Can I do 25 or 90?

Forty-five minutes is a default, not a prescription. It sits within the natural ultradian rhythm cycle described in sleep research—approximately 90-minute rest-activity cycles—capturing the ascending alertness arc without pushing into the declining phase.

If you are new to structured focus, 25 minutes is a reasonable starting point. The protocol works the same; the block is just shorter. For deep analytical or creative work where you need longer immersion, 60 to 75 minutes can work well once the queue-offloading mechanism is established.

What does not work is extending blocks because you do not want to stop. Block length should be set before the session, not adjusted mid-session. If you want a longer block, set it at the start.

Q: Do I really need the five-minute setup, or can I just start working?

You can start without it. Many people do. The question is whether you want the structural benefit or just the time block.

The setup prompt does three things: it defines the task precisely (which reduces initiation friction and mid-session ambiguity), it offloads the queue (which removes the primary source of attention residue during the session), and it recruits the AI as a focus partner for the session. Skipping it means skipping those benefits.

The five minutes is overhead in the same way a pre-flight checklist is overhead. It takes time; it also prevents avoidable failures.

Q: What happens if I skip the unlock?

You exit the session with the queue unsurfaced, captures not integrated, and the transition unstructured. The next activity gets whatever attention is left after your brain attempts to reconstruct the queue without the AI’s help.

Most people who skip the unlock report that the next thirty to sixty minutes are less productive than sessions where they ran it. The unlock is not just a review—it is the cognitive closure mechanism that lets you fully release the previous session.

Practical Challenges

Q: My work involves a lot of communication. How do I single-task when responding quickly is part of my job?

The protocol is not about eliminating responsiveness. It is about batching it.

During a 45-minute lock-in, communication goes to queue. When the session ends, you process the queue as a deliberate activity—which is itself a single-tasking session. Communication work, done in concentrated batches with full attention, is typically faster and higher quality than communication done reactively in parallel with other work.

The practical step: discuss response-time expectations with your team. Most people overestimate how quickly responses need to arrive. Establishing that a 45-minute response window is acceptable for most communications removes the anxiety that drives reactive checking.

Q: What do I do when something genuinely urgent arrives mid-session?

Break the session. Handle the urgent item. Restart.

When you restart, run the setup prompt again—even if you were only twenty minutes into the previous block. The re-setup primes the correct cognitive mode and re-establishes the queue. Do not try to resume where you left off without it.

Define “urgent” before the session begins. Write it down. “Urgent means: [specific manager] with [specific escalation flag] only.” Everything that does not meet that definition goes to the queue regardless of how urgent it feels. The definition protects you from your own urgency bias.

Q: I tried single-tasking before and found it made me anxious rather than focused. Why does that happen?

The anxiety is usually queue anxiety—the sense that something important is being missed while you are locked in.

This is the primary problem that AI-assisted queue offloading solves. The anxiety is not irrational; it arises because, without an external holding system, the queue is genuinely at risk of being lost. The solution is not to will yourself to be less anxious. It is to address the underlying structural cause by giving the queue to a trusted system.

If you have tried the protocol with the AI queue offloading and still feel anxious, two further steps: (1) check whether your definition of urgent interruptions is clear—if it is vague, any incoming item will feel potentially P0; (2) run shorter sessions initially (25 minutes) until the AI queue system has proven itself reliable through experience.

Q: My colleagues see me sitting quietly on one task and assume I am not busy. How do I manage the social dimension?

This is a real problem in cultures that reward visible busyness.

Two approaches work together: transparency and outcomes.

Transparency means telling your team what you are doing. “I do a 45-minute focus block most mornings—I’m unreachable during that window except for X. Here’s how to reach me for X.” Most reasonable colleagues will accept this once it is explicit.

Outcomes means letting your output quality speak. Over time, the work you produce from structured focus sessions tends to be more complete, more accurate, and more useful than work produced reactively. That quality difference becomes visible. It builds the credibility that sustains the practice against cultural friction.

Q: Can single-tasking work for creative work, where ideas come from unexpected connections?

Yes, with task definition adapted for exploration.

Output-defined tasks work for creative work—the output is just more exploratory. “Generate three distinct directions for [project]” is a valid task for a 45-minute lock. “Explore the design space freely” is not—it lacks a completion criterion and will drift.

The key is that creative exploration still benefits from uninterrupted time. A 45-minute lock with an exploratory task definition produces more interesting creative output than 45 fragmented minutes of switching between the creative task and other inputs. The creative mind still benefits from sustained immersion; it just needs a somewhat looser task frame.

Getting Started

Q: I want to try this but feel like I do not have forty-five uninterrupted minutes in my day. What then?

Look at your actual calendar, not your ideal calendar. Most people have at least two to three windows that are nominally uninterrupted but currently used for reactive work. A fifteen-minute gap between meetings, a pre-meeting window, the first or last hour of the day.

Start with twenty-five minutes. Any gap you currently fill reactively is a candidate for a mini-lock. Run the protocol at the shorter duration. The benefits scale down accordingly but they are present.

Q: What is the single best thing to do to get started today?

Copy the setup prompt from the 5 AI Prompts for Single-Tasking article, fill in the brackets with your most important current task and the items you are mentally holding, and run one session before the end of today.

Do not optimize the protocol before you try it. Run it once, note what felt different, and adjust from there.

Related:

• The Complete Guide to Single-Tasking with AI Support
• How to Single-Task with AI Support
• The Science of Single-Tasking
• Why Multitasking Feels Productive

Tags: single-tasking, FAQ, AI productivity, attention management, One Thing Lock