Picture the finished future, then keep asking “what had to happen just before?” until you're back at today.
Forward planning asks “what should we do first?” and often stalls, because from today there are endless possible next steps. Reverse Planning flips the direction. You describe the future as if it has already happened, then ask again and again, “What had to be true just before that?”, until you arrive back at today.
Starting at the end has two big benefits. It lets you aim high: you can begin with an ideal, even slightly impossible, picture and pull it back step by step to a version you can actually reach, so the ambition survives contact with reality. And it shows which steps are unavoidable and how much time each really needs, so deadlines stop being a surprise.
Use it when you have a fixed date or a bold goal, when forward plans keep running out of time, or when an exciting idea seems impossible and you want to find the achievable version hiding inside it.
Start at opening night and keep asking what had to happen just before, until you reach a step you can take today.
If this went better than you dared hope, what would the finished future look like?
Which version of that future keeps the heart of the dream but could really be done by your date?
For the finished future to be true, what had to be true just before it? And just before that?
Keep stepping back: what is the very first thing that has to happen, and can it start this week?
Read the plan from today forward: does it flow, who owns each step, and where is the slack?
Starting at the end changes the question you ask. Instead of “what could we do?” you ask “what must already be true?”, and that changes what you notice.
From today, almost anything could be a next step, which is why forward plans drift. From the finish, the question is narrower: what must already be in place? Usually only a few things qualify, so the route shows up fast.
Most bold ideas die in the first minute, killed by “but how would we ever…?” Here the how waits until the picture is clear. Then you trade impossible parts for possible ones, one at a time, and keep more ambition than a cautious plan would.
Working back from a fixed date shows how much time each step can have. You discover the long lead times, such as approvals, bookings and deliveries, while there's still time to start them, not the week before.
Because each step exists to make the next one possible, nothing is in the plan out of habit. If a task doesn't lead to a later milestone, it's a candidate to cut.
Describe the end as if it has already happened, in the present tense: who is there, what's different, how it feels. Let it be ambitious, even a little impossible, for now.
Step from the ideal toward reality. Keep the heart of the dream but swap impossible parts for achievable ones. Write a checkable “done” statement with a date.
From the finish, ask what had to be true just before it, then just before that. Write each milestone on a card, moving back toward today, with how long it really takes.
Keep stepping back until you reach something you can start this week. If the steps don't fit between today and the deadline, change the scope, the date or the resources now.
Read the plan from today to the finish. Does it flow? Add owners and dates, and build in buffer time for the steps most likely to slip.
Compare progress with the milestones. When something slips, plan backwards again from the finish to find the new route.
A long wall or sheet with the finish at one end and today at the other, sticky notes and a calendar. With AI, the prompt below can walk you back one step at a time.
Who: a university's student-success office. Finish: about 400 first-years matched with trained student mentors in week one of fall term. Constraint: a small budget, and the student planning team leaves campus in May.
Every new student has a mentor who feels like a friend, nobody slips through the cracks, and mentors call it the highlight of their year.
Week one of fall: 400 first-years matched with 80 trained mentors, with each group's first meeting already on the calendar.
Matches sent and first meetings booked. That needs the list of incoming students from admissions by late July.
Admissions shares the list once the privacy office signs off. Approvals can take weeks, so the request goes in now.
Mentors trained on campus. It has to be April: once students scatter for the summer, a training day is almost impossible.
80 mentors recruited, with a two-minute application and a nudge from professors in large first-year courses.
This week: send the data request, book a room for the April training and open the mentor application.
Planning forwards, the team would have trained mentors in August, after most had left campus. Planning backwards caught it in time.
Two weeks of slack before the April training, a backup plan to train late recruits online in August, and a progress check on the first Monday of every month, with a fresh backwards pass whenever a milestone slips.
Reverse Planning asks you to imagine a future in detail, then reason backwards one prerequisite at a time. AI is good at both: painting the finished picture, asking “what had to happen first?” and spotting the lead times people forget. You know the real deadlines, people and politics, and you decide what's worth the effort.
Start bold, then step back toward a version you could actually reach.
Describe the ideal version of our mentoring program. Then give a version we could really reach by fall.
Let AI keep asking for prerequisites so that none get skipped.
Finish: 400 students matched by fall. Keep asking me what had to happen just before, until we reach today.
Find the slow steps that must start long before they feel urgent.
Which steps in my plan have long lead times, like approvals, bookings or hiring? When must each start?
Imagine the plan failed, then work out which milestone slipped first.
It's fall and the program failed to launch. Which milestone most likely slipped first, and why?
Describe your goal and deadline, pick a mode, then copy the prompt into Claude, ChatGPT, Gemini or Copilot. In coach mode the AI starts at the finish with you and walks back one step at a time, waiting for your answers.
The “Open in” links put the prompt in the web address. If your challenge includes anything sensitive, copy and paste instead.
After the session, click one to copy it, then paste it into the same chat.