Planning With AI: Put the Real Obstacle Next to the Goal

Imagine an independent illustrator asking AI to help launch a weekly email. Within minutes there is a name, a six-week calendar and a welcome message. The plan looks ready to go. But the illustrator’s only quiet editing time is on Friday, and client revisions regularly take it. The calendar explains what a successful newsletter would look like. It does not explain what will happen on the first Friday that runs late.

An AI assistant can make an attractive future easy to picture. Before you commit to its timetable, put that future beside one real obstacle. You may discover a practical change that makes the goal worth pursuing. You may also discover that the proposed version asks for time, authority or resources you do not have.

What mental contrasting adds

In a May 2001 paper on goal setting, Gabriele Oettingen, Hyeon-ju Pak and Karoline Schnetter tested a technique now commonly called mental contrasting. Participants considered a desired future and then reflected on the present reality that stood in its way. Across four experiments, commitment in the contrasting conditions depended on expectations of success: favourable expectations were associated with stronger commitment and unfavourable expectations with weaker commitment.

The finding is useful because it leaves room for a goal to become less compelling after you examine it. The studies did not test AI planning assistants, and they do not show that a chatbot can estimate your prospects. The practical application here is to use the conversation to make the contrast concrete, then judge the evidence yourself. A model can offer possibilities; it cannot observe your Friday unless you tell it what happened.

Describe the result and the obstacle separately

For the illustrator, the desired outcome might be that readers receive one useful sketching exercise each week. That is more specific than ‘build a newsletter’, and it may be easier to deliver than an issue with a feature essay, a video and a resource list. Ask what the reader should be able to do after opening it. Keep that result visible while you consider how to produce it.

Then pick a recurring situation that has already blocked similar work. ‘I get busy’ is too broad. ‘Client corrections arrive on Thursday and take the only editing block on Friday’ gives you something to check. Look back at the calendar or job notes. Was Friday the real bottleneck, or was it the only day you tried? If the issue is waiting for a collaborator’s approval, protecting Friday may make no difference. A good obstacle description points to what would need to change.

Resist the urge to ask for a long list of ways to be more productive. Ask the AI to repeat the constraint in plain language and suggest two different responses, explaining what each assumes. One response might move preparation earlier in the week. Another might shrink the first issue to an annotated sketch the illustrator already has permission to share. If the assistant proposes Saturday work, tell it whether Saturday is actually available. A calendar cannot solve a capacity problem by allocating time that does not exist.

Use a small attempt to check feasibility

Suppose the illustrator can protect half an hour on Wednesday. Use it to assemble one sample issue from an existing sketch. Check what can be finished within that time and what still needs attention. If the writing fits but image preparation takes another hour, the trial has exposed a better obstacle than ‘not enough discipline’. The next version could use a simpler layout, or the frequency could change. The test should teach you something even if you decide not to launch.

Distinguish low confidence from a practical constraint. Feeling unsure about a first issue might call for a private trial with a trusted reader. Lacking permission to share the underlying images calls for a different source of material or a delay. The AI can help outline either response. It should not turn a missing permission or budget into a pep talk. If you do find a change that makes the goal feasible, our guide to building an if–then plan can help you decide when the action starts.

If the goal depends on another person, turn the assumption into a question for them. A collaborator may be happy to approve one sketch a month but unable to review one every Thursday. Ask before designing a schedule around their availability. Their answer might rule out the weekly version while leaving a monthly issue entirely feasible.

Give the planning chat a clear job

• Name the outcome in terms someone else could observe. Say who benefits and what the smallest useful version would be.
• Describe one recent occasion when the obstacle showed up. Include what happened, when it happened and which part of the work it stopped.
• Ask for two responses that address that obstacle directly. Request the assumptions behind each and discard any that depend on resources you cannot use.
• Choose a small trial and agree what would make you continue, scale down or wait. Record what happened before asking AI to revise the plan.

After the trial, compare your expectations with the work you actually did. If the smaller issue takes less time and still helps a reader practise, you have evidence for a version you can repeat. If the obstacle holds, you have a reason to change the goal or its timing. Either outcome is more useful than another polished calendar built on the same unavailable Friday.

Source

• Oettingen, Pak and Schnetter, Self-Regulation of Goal Setting: Turning Free Fantasies About the Future Into Binding Goals (2001)