How to Plan a Project with AI: A Prompt Workflow

A single-location coffee shop deciding to open a second site runs straight into a planning wall. The lease, the build-out, the permits, the equipment, the hiring, and the supply chain all move on different clocks, and the owner is still working the morning bar shift. Two months of scribbled notes later, nobody can say what must happen before what.

This workflow uses the model as a project planner you interview, not an oracle you trust. You provide the constraints and the budget; it turns them into workstreams, dependencies, owners, and a risk register you can act on before the first contractor is called.

The Output You Are Building Toward

A one-page project plan with five workstreams, a milestone timeline, a raci-style owner list, a budget line, and a ranked risk register. It fits on a single sheet, which matters when the people executing it are also pulling espresso shots.

Step 1: State the Goal and the Constraints

Models plan badly when the goal is vague. Force a crisp scope, a hard budget, and the non-negotiables before any task appears.

You are a project planner for a small business. Help me plan opening a second location for a single-site specialty coffee shop.
Goal: open to the public within 6 months.
Budget: 120,000 USD total, including build-out and first 3 months of rent.
Constraints: the owner still works 30 hours a week at location one; only one full-time manager can be hired before opening; the lease is not signed yet.
Before you plan, ask me up to 4 questions that would change the plan most.

You get a short interview that surfaces what you forgot, such as whether the new site shares a roastery or needs its own. Answering those four questions now prevents a plan that quietly assumes the wrong supply chain.

Step 2: Break the Work into Workstreams

A flat task list hides the work that runs in parallel. Group the effort into workstreams so you can see what can proceed at the same time.

Organise the project into workstreams. For a second coffee shop these likely include site and lease, build-out and permits, equipment and suppliers, people and training, and launch and marketing.
For each workstream, list 5 to 8 concrete tasks. Mark each task as a milestone (a decision or a deliverable) or a routine task.

You get five labelled columns of work instead of one intimidating list. Milestones like “lease signed” and “health permit approved” become the checkpoints the whole plan hangs on.

Step 3: Map the Dependencies and Timeline

The real reason plans slip is a task scheduled before something it depends on. Extract those relationships explicitly.

For every task in each workstream, state what it depends on. Then produce a timeline in weeks (week 1 to week 26) that respects those dependencies.
Call out any task that sits on the critical path, meaning the whole opening date moves if it slips. Assume permits take 4 to 8 weeks and build-out takes 8 to 12 weeks.

You get a week-by-week sequence where the critical path is named. In a coffee shop that path usually runs lease, then permits, then build-out, then equipment install, then training, then soft launch.

Step 4: Assign Owners and a Budget Line

Tasks without owners do not happen. Attach a name and a rough cost to each workstream so nothing hides in the gap between “the owner will handle it”.

Assign an owner to each workstream from this team: owner, new store manager, a part-time bookkeeper, and an external contractor.
Then allocate the 120,000 USD budget across the five workstreams, and flag which workstreams are most likely to overrun.
Keep a 10 percent contingency unallocated.

You get an owner column and a budget table where the contingency is visible. Seeing that build-out and permits carry the highest overrun risk is usually enough to make you negotiate the lease harder before signing.

Step 5: Build a Risk Register

Every second location has three or four ways to fail. Name them early and attach a response.

Create a risk register with columns: Risk, Likelihood (low, medium, high), Impact (low, medium, high), Mitigation, and Owner.
Include risks specific to a second coffee shop site: construction delays, permit rejection, hiring delays for trained baristas, and supply disruption if the new site depends on the same roastery.
Rank the register by Likelihood times Impact.

You get a ranked list where permit rejection and hiring delays typically top the table, each with a pre-agreed response. That turns a crisis into a decision you already rehearsed.

Step 6: Compress It to One Page

The plan is only useful if the team reads it. Force a summary that fits one screen.

Condense everything above into a one-page project brief: goal, five workstreams with owners, the 26-week milestone timeline, the budget allocation, and the top 5 risks with mitigations.
Use plain language. No jargon. If the brief runs past one page, cut routine tasks and keep only milestones.

You get a single sheet the owner and the new manager can both hold. Routine tasks live in a backlog; the one pager tracks the milestones that decide whether the doors open on time.

Final Deliverable Checklist

• A goal with a hard deadline and a fixed budget.
• Five workstreams, each with five to eight tasks.
• A 26-week timeline with the critical path named.
• An owner for every workstream, including the part-time roles.
• A budget allocation with a visible 10 percent contingency.
• A ranked risk register with a named response per risk.
• A one-page brief the whole team can read.

Common Pitfalls

Pitfall 1 — accepting a plan with no contingency. If the model allocates the full budget with no reserve, ask again. A second location almost always surprises you on permits or build-out.

Pitfall 2 — planning the opening as a single task. Opening day is the end of a chain: training, first stock order, POS setup, and a soft launch for friends. Break it out or it will be rushed.

Pitfall 3 — forgetting the day job. The owner still works the bar. If the plan assumes forty hours a week of owner time, it is already broken.

Pitfall 4 — treating the timeline as fixed. Re-run the dependency step whenever the lease date moves. Everything downstream shifts with it.

To keep the plan alive week to week, pair it with our weekly work plan prompt. For the meetings that will follow, see meeting notes summary prompts, and browse more systems in our productivity prompts list.

Worked Example: The Critical Path

Here is what the dependency step looks like once it is done, so you can see the plan is a sequence rather than a pile.

Week 1-4: negotiate and sign lease (critical path)
Week 5-12: permits submitted and approved (critical path, 4-8 week window)
Week 6-16: build-out, runs partly in parallel with permits (critical path)
Week 17-19: equipment delivery and install (critical path)
Week 20-22: manager hired and baristas trained (critical path)
Week 23-24: first stock order and POS live
Week 25: soft launch for staff and friends
Week 26: public opening

Everything off the critical path — marketing, menu design, loyalty program setup — can flex into the buffer weeks without moving the opening date.

Quick FAQ

Should I plan for six months or twelve? Plan the six months to opening in detail, then sketch the first ninety days of trading separately. The opening plan and the operating plan are different problems.

What if the budget is fixed and the plan exceeds it? Ask the model to re-sequence by value: which tasks protect the opening date, and which can be deferred to month two of trading. A second espresso machine can wait; a permit cannot.

Can I use this for a project that is not a physical space? Yes. Swap the workstreams for your context and keep the dependency and risk steps, which are what make the method work.

How often should I re-run the plan? Revisit it every week against the milestones, and fully rebuild the timeline only when a critical-path date moves.

Sources & Notes

Workstream planning, dependency mapping, and risk registers are standard project-management practices; the prompts and structure here are our own, reviewed on 2026-10-09. Budget figures and timelines are illustrative starting points — replace them with real quotes from your contractor, landlord, and equipment supplier before committing money.