A discovery phase turns a rough idea into a plan a team can build and a client can sign off. This project discovery phase template is a discovery phase document template: it lists the ten sections a good discovery document covers, what to write in each and how to tell it is finished. Copy the headings into your own document, or let Discovery Phase AI draft them from your existing decks, notes and call transcripts.
1. How to use this template
The template works as a project discovery phase document, also called a discovery report: one file with ten sections instead of ten separate files. A single document is easier to read, to review with the client and to keep current.
- Copy the ten headings below into your own document (Word, Google Docs, Notion or Markdown all work).
- Fill the sections in this order: stakeholders and background, problem statement, goals, users and journeys, requirements, architecture, UI screens, then the roadmap, estimates and risks. Write the executive summary last.
- Mark anything you do not know yet as an open question with an owner, instead of leaving the section empty or guessing.
- Check each section against its "done when" line before you share the document.
- Have the sponsor read the executive summary and agree the next step: build, build something smaller, or stop.
2. Stakeholders
Name everyone who can approve, block or change the project. The goal is that nobody is surprised late.
- Name, role and organization
- What they need from the project
- Influence (decides, advises, informed) and how to reach them
- Open disagreements between stakeholders
Done when: you know who signs off and who is allowed to say no.
3. Background
The context a new team member needs on day one: the business, the current product or process, past attempts and anything that constrains the work.
- Company and market in a few sentences
- How the work is done today and with which tools
- Previous projects, what worked and what did not
- Hard constraints: budget, deadline, regulation, existing systems
4. Executive summary
One page for the person who will not read the rest. Write it last, but keep it short enough to be read in two minutes.
- The problem and who has it
- The proposed solution in plain words
- Expected outcome and how it will be measured
- Scope, rough timeline and cost range
- The decision you need from the reader
5. Problem statement
Describe the problem, not the solution. A useful statement says who is affected, what goes wrong, how often and what it costs.
Template: [Who] cannot [do what] because [cause], which leads to [measurable impact]. We know this because [evidence].
6. Goals and success metrics
Two or three goals, each with a number and a date. A goal without a metric cannot be judged later.
- Goal: what changes for the business or the user
- Metric: how it is measured, with today’s value and the target
- KPI: the figure you will keep watching after launch
7. Users, journeys and flows
Show who uses the product and what they do with it.
- Users and actors, with their goals
- Journeys: the steps a user takes to reach a goal, including where they get stuck today
- Flows: key processes as simple diagrams
8. Requirements
What the product must do, written so that each item can be tested.
- A short title and a one-sentence description
- Type: functional, non-functional or constraint
- Priority: Must, Should, Could or Won’t (MoSCoW)
- Acceptance criteria: how you will know it works
- The goal it supports and an owner
Done when: every Must has acceptance criteria and traces back to a goal.
9. Architecture and decisions
A first technical picture, and a record of why you chose it.
- A context diagram and a component diagram
- Integrations and data that cross system boundaries
- Decision records: the question, the options compared, the choice and its consequences
10. UI screens
Rough wireframes of the screens that matter most. They are for agreement, not for design polish.
- The three to five screens users meet most
- Which requirements each screen covers
- Open questions the wireframes raised
11. Roadmap, estimates and risks
Milestones with a range, never a single number, plus what you assumed and what could go wrong.
- Milestones with start, target date and a low–high effort estimate in person-weeks
- Dependencies between milestones
- Assumptions that the estimate relies on
- Risks, blockers and open questions, each with an owner and a next step
12. Example: a filled-in problem statement and goal
The example below is invented, from a made-up field-service app called Fieldwise. It shows the level of detail to aim for.
Problem statement: Dispatchers cannot see that a technician has declined a visit until the technician phones in, often hours later, which leads to about one visit in ten being missed. We know this because the dispatch team logged every missed visit for a month.
Goal: cut missed visits from about 10% to under 3% within six months of launch. Metric: missed visits as a share of scheduled visits. KPI: the same figure, reviewed weekly by the head of dispatch.
Note what the statement leaves out: it names no app, screen or technology. The solution comes later, in the requirements and the architecture.
13. How much of the template to use
These are rules of thumb. Adjust them to your risk and your budget.
- Small MVP: go deep on the problem statement, the goals and the Must requirements. Keep the architecture and UI screens short.
- Typical product: go deep on requirements, integrations and the roadmap with ranges. Keep visual design out of discovery.
- Regulated or legacy-heavy project: go deep on non-functional requirements, integrations, data and compliance constraints, and risks. Skipping a section here costs the most later.