- Foundation
- 30 min
- 5 min read
After this lesson you can
- Name the three parts of assessment: checkpoint, rubric, final review.
- State the pass level and the rule about skills below level 2.
- Say what the final panel checks.
- Explain what a coaching plan is and why it is not a failure.
Read these first
Assessment in this program is not a surprise test. It is a steady conversation about work you can show, with three parts and one clear pass level.
Why it matters
When you know how you will be judged, you can aim at it. The program does not grade your polish. It grades whether the thinking is yours, whether it covers the whole product, and whether agents can build from it.
The three parts
| Part | What it is | When | Time |
|---|---|---|---|
| Weekly checkpoint | You show your work and your Sensei asks questions. You do not need to be perfect. You need to be honest and improving. | Every week | About 30 minutes |
| Self-assessment rubric | You score yourself each week on six skills. Your Sensei scores you too. You compare. | Every Friday | 20 minutes |
| Final roadmap review | You present your roadmap and defend it to a panel | End of Week 4 | A ten-minute presentation, then questions |
The rubric pages are Track A self-assessment for the planning level, Track B gates and the capstone rubric for the platform level. The weekly pages end with a self-check you can use at once.
What the panel checks
- Is it your thinking? Can you explain every epic without notes?
- Does it cover the whole product? Not only the app: the site, the services, the money, the people.
- Can agents build from it? Are the milestones clear enough that an agent has few questions?
The pass level
At the end of Week 4 you should reach level 3 (Competent) or higher on all six skills, and no skill below level 2.
| Level | Name | Meaning |
|---|---|---|
| 1 | Beginning | I am starting and need help with the basics |
| 2 | Developing | I can do part of it, with gaps |
| 3 | Competent | I can do it well on my own, with evidence |
| 4 | Strong | I can do it well, predict problems and help others |
Score one level lower if you cannot point to evidence in git.
If you fall short
Falling short means a coaching plan, not a failure. A coaching plan is a short list of actions with dates, agreed with your Sensei. You work through it and review again. Nobody is removed for needing more time.
The platform level
The Master Developer level uses four gates, a final exam, a practical capstone and an oral review. Pass mark: 70 points and all mandatory items. Five critical behaviours fail a practical at once: no secrets in git, no direct changes to a cluster, fix forward, check the licence, and the merge request is the change record. See Master Developer assessment.
Steps for each week
- On Friday, score the six skills. Write one piece of evidence for each score, linked in git.
- Write your weekly note, with the AI use line.
- Bring the score sheet to your checkpoint.
- Compare scores. Ask what your Sensei saw that you missed.
- Agree one thing to fix.
Working with AI agents
- May: check that each score has a link to evidence; calculate your average per skill over the weeks.
- May not: score you, write your evidence, or fill the rubric for you.
- A safe brief: “For each row of my score sheet, say whether it has a link to evidence. List rows with none. Do not change scores.”
Common mistakes
- Scoring yourself high with no evidence.
- Arguing about the number. Ask about the gap.
- Treating a coaching plan as shame.
- Preparing the slides and not the answers.
- Hiding AI use. It is worse than the use.
Practice
- Score yourself on one skill today and write the evidence link.
- List three questions the panel could ask about your roadmap.
- Write a one-line coaching plan for your weakest skill.
Checklist
- I know the three parts of assessment.
- I know the pass level.
- My scores have evidence links.
- I know what a coaching plan is.
The 2030 Product Development Manager lens
- You decide: your honest score and what you will fix.
- You delegate to agents: evidence-link checks and score arithmetic.
- A sample brief: “List every evidence link in my score sheet that does not open. Do not edit.”
- You verify by: opening three links and reading what they show.
- Disciplines involved: quality, product and coaching.
- On the agent ladder: assessment is how people govern: a human check of work that agents helped to make.
Further learning
Next: the career path. Read the weekly checklist.
Go deeper
Further learning
Free or low-cost material from other publishers. We opened every link; check the cost label before you start.
DORA
DORA: Get Better at Getting Better (opens the publisher's site in a new tab)
- Free
- Docs
- Beginner
The research programme behind the well-known delivery metrics, with guides and capabilities that help a team improve. Use it to see delivery as a set of habits you can measure and improve.
GOV.UK Service Manual
Agile delivery (opens the publisher's site in a new tab)
- Free
- Docs
- Beginner
The UK government's guide to agile delivery for public services: teams, phases, user stories and how to work in short cycles. Written in plain English, which also makes it good for learners with English as a second language.
Atlassian
What is the Definition of Done (DoD) in Agile? (opens the publisher's site in a new tab)
- Free
- Article
- Beginner
Explains a shared checklist that says when work is truly finished. It is the human version of the acceptance checks you give an AI agent.
Atlassian
How to write SMART goals (with examples) (opens the publisher's site in a new tab)
- Free
- Article
- Beginner
A short article that explains the five SMART tests (specific, measurable, achievable, relevant, time-bound) and works through an example. Use it when you rewrite a vague task into one that an agent or a colleague can finish and you can check.
Atlassian
What are agile retrospectives? (opens the publisher's site in a new tab)
- Free
- Article
- Beginner
Describes how a team looks back at its last stretch of work and agrees what to change. A practical guide for running your own weekly review, alone or with agents in the loop.



