AI Planning: Pair Enjoyment with a Routine Task

The empty mailing boxes are still stacked beside the desk. You know they need folding before tomorrow’s orders, but the task keeps losing to something more appealing. An AI assistant could produce a stricter timetable. Another possibility is to ask whether the task could accompany something you already want to enjoy.

That question connects to temptation bundling: pairing an immediately enjoyable activity with a useful activity whose rewards arrive later. Research offers evidence for particular pairings in particular settings. For a creator’s working day, the interesting challenge is to design a modest trial without turning the research into a universal productivity promise.

What the original experiment tested

Katherine Milkman, Julia Minson and Kevin Volpp examined this idea in Holding the Hunger Games Hostage at the Gym, published online on 6 November 2013 and in the February 2014 issue of Management Science. Participants were randomly assigned to gym-only access to tempting audiobooks, encouragement to restrict their audiobook listening to the gym, or a control condition.

The authors reported initial increases in gym attendance in the treatment groups, with effects declining over time, particularly after Thanksgiving. The comparison involved a specific activity, an appealing audio experience and different arrangements for access. It was not a test of adding entertainment to every kind of work.

The distinction between an imposed restriction and a suggested personal rule also matters. A planner that offers an optional pairing is making a different intervention from a system that controls when someone can access the enjoyable activity.

Later research separated encouragement from access

A November 2020 paper by Erika Kirgios and colleagues studied participants in a four-week exercise programme. It compared providing an audiobook with encouragement to bundle, providing an audiobook alone and relevant control conditions. The authors reported benefits for gym visits, including during a seventeen-week follow-up.

The added benefit of explicit bundling encouragement over an audiobook alone was modest. Encouragement alone did not robustly increase the number of weekly visits, even where it increased the likelihood of a visit. Receiving enjoyable content and receiving an instruction about how to use it should therefore remain separate explanations.

These studies concern attendance. They do not establish that the same arrangement improves creative quality, produces accurate administration or makes AI planning effective. Applying the idea to a studio routine is an editorial proposal that needs its own practical check.

A fictional studio pairing makes the choice concrete

Imagine an independent printmaker who needs to fold twenty empty mailing boxes. They would also like to continue a favourite audiobook. The proposed pairing is to listen while preparing that small batch, then stop and check whether the boxes are correctly assembled.

This is different from promising an audiobook after the entire day’s work. The enjoyable activity accompanies the task. It is also different from asking AI to fold the boxes: the assistant’s proposed role is to help specify the arrangement, while the printmaker chooses and performs it.

Start by naming the actual obstacle. If the boxes are missing, the instructions are unclear or there is no usable workspace, the immediate job is to resolve that problem. Pairing entertainment with an undefined task would leave the practical gap intact.

In this fictional example, the materials and method are already available. The open question is whether the printmaker likes doing this particular routine alongside the chosen audio. Neither the studies nor an assistant’s confidence can answer that personal question in advance.

Ask AI for compatible options, then choose one

A bounded planning brief could state the task, the available time, the materials and the enjoyable activity the person already prefers. Ask the assistant for a small number of possible pairings and the checks each would require. The useful output is a clear choice, rather than an elaborate schedule for every minute.

• Task: prepare one batch of twenty empty mailing boxes using the established method.
• Possible accompaniment: the next part of an audiobook the printmaker already enjoys.
• Completion check: count the assembled boxes and inspect whether they are usable.
• Review: decide whether to keep, change or abandon the pairing after the session.

The same printmaker might decline audio while checking postal addresses or proofreading a product description. That is a task-design choice for this example, not a finding from the cited experiments about those jobs. Keep the person doing the work involved in deciding whether an accompaniment fits.

If an assistant proposes constant conversation, notifications or another activity instead, assess that proposal separately. The research on gym visits with audiobooks does not validate every form of simultaneous stimulation. Ask what the proposed arrangement contributes to the actual task.

Review the work as well as the willingness to begin

Suppose the printmaker finishes the session and reports that it felt more inviting. That observation is worth recording alongside the output: how many boxes were completed, whether any needed redoing and whether the session fitted the time available. An enjoyable experience and a usable batch are two relevant outcomes.

One session cannot establish that the pairing caused a lasting change. The task, workload and mood may differ next time. A brief personal note can still help the printmaker decide what to try again without presenting the experience as a research result.

If the audiobook became the main attraction and the boxes remained unfinished, change the arrangement. If the work went well but the person disliked the restriction, reconsider the rule. The purpose is a workable routine chosen by the person, rather than loyalty to a technique.

Keep the experiment smaller than the job

There is little value in spending an hour asking an assistant to optimise a short packing task. For this example, finish the planning once the batch, accompaniment and review are clear. Leave room to perform the work and discover whether the pairing is useful.

MaryChuks.com’s article on how the next task feels after AI help offers a related perspective: judge a workflow through its result and the person’s experience. Here, the added question is whether enjoyment belongs alongside this particular routine.

Temptation-bundling research gives a reason to consider how immediate enjoyment and useful action can coexist. It does not remove the need to inspect the task or hear the person’s preference. Let AI help frame one small option, then let the quality of the work and the experience of doing it guide the next decision.

Sources

Primary sources checked on 30 October 2026. The printmaker, box-folding task and AI planning brief are fictional applications. Neither cited paper tests AI-assisted planning or studio productivity.

• Milkman, Minson and Volpp: Holding the Hunger Games Hostage at the Gym: An Evaluation of Temptation Bundling — Management Science; online 6 November 2013, February 2014 issue.
• Kirgios and colleagues: Teaching temptation bundling to boost exercise: A field experiment — Organizational Behavior and Human Decision Processes, November 2020. Author-hosted paper.