Mentor · for the Product Manager

Your field guide

You're running Mentor's product queue. This is the whole job, in pictures — how to read what testers ask for, decide what gets built, keep everyone posted, and stay inside a budget. No code. Read it once, then keep it open on your first pass through the console.

🔑 Getting in: open mentor.arnao.ai, log in with Google (your usual account). If you were already open in the app, refresh the page once. Then tap the ⋮ menu (top-right) → Switch role → Product.

1Find your console

Tapping ⋮ → Switch role shows you two doors. Mentee is you as a normal user (your own goal & plan). Product is your operator console. You won't see Mentor or Admin — those aren't yours, and that's on purpose.

Switch role menu showing Mentee and Product — The role switcher — tap **Product**.

2Read the room

This is your home base. A quick tour of what you're looking at:

The budget strip (very top): your live spend and your cap. When it's used up, the app's AI pauses until Byron tops it up — that's the guardrail, not a bug.
The banner: the one rule to remember — you handle low & medium risk; high & unclassified go to Byron automatically.
The tiles: queue size, what's pending, what's approved and ready, open bugs, releases, and the live app version.
The queue: every piece of tester feedback, newest triaged first, each tagged with a risk pill.

The PM console backlog with risk pills — Your console: budget strip, guardrail banner, tiles, and the triaged queue.

3The risk pills — your traffic light

An assistant reads every incoming item and tags it. You act on green and yellow; red and grey are Byron's.

RISK·LOWCopy tweaks, styling, small polish. Yours — approve freely.

RISK·MEDChanges to how an existing screen behaves, new AI wording. Yours — approve with a moment's thought.

RISK·HIGHLogin, accounts, emails, anything that could break onboarding. Byron's — the app will stop you if you try.

RISK·?Not yet sorted by the assistant. Treated as Byron's until it's classified.

Try it: tap Approve on a red HIGH item on purpose. It should refuse with "Needs Byron." If it ever lets you, that's a real bug — tell Byron.

4Approve → bundle → Create PR

Approving an item says "yes, this is worth doing." Building happens by bundling approved items into a release — that's what "Create PR" means (PR = the work order that gets built).

Tap Approve on the items you want.
Tick the checkbox on two or more approved items.
Hit ⚡ Create PR from selected. That writes the build brief and notifies the build side. You never touch code.

Two approved items selected with the Create PR bar — Two approved items selected → the **Create PR** bar lights up.

5Tell the testers — Beta update

When a release ships, keep the testers in the loop. In the Releases list, each release has a 📣 Beta update button. Write two or three sentences — what changed, what to try.

It sends a test copy to Byron first, on purpose, so nothing goes out unseen.
Real sends reach every active tester and are capped at 4 a month.

Releases list with the Beta update button — Each release carries a **Beta update** button.

6Your budget

💷 You have a $15 weekly working budget. The app enforces it — you can't overspend. If you hit it, AI features pause until Byron refreshes it. Managing inside the budget is part of the role: it's a real constraint, on purpose.

Something bigger or more expensive you want to build? That's exactly when to ask Byron rather than push it through — anything costly is his call. More on how building works below.

7How a feature actually gets built

You don't write code, and neither does Byron by hand. Here's the chain your "Create PR" kicks off:

You approve + Create PR→ Byron reviews the plan→ a build assistant writes it→ preview→ Byron promotes it live

Small, cheap work gets built by a low-cost assistant; anything expensive or risky pauses for Byron's OK. You stay focused on what to build and why — the machinery handles the how.

8Now the real ask — what should we build?

Reading testers is half of it. The other half is your judgement. As you go, keep a running note on:

What would you actually use? You're a designer and a user — the features you'd reach for are probably the ones others want too.
What keeps coming up? If three testers trip on the same thing, that's a signal, not noise.
What would make someone come back tomorrow? Not just fixes — the one thing that makes Mentor feel alive.
What's confusing on first use? You'll only have fresh eyes once. Write it down now.

Bring those to Byron. The queue is where testers talk; you're the one who decides what it means.

On your phone

Everything works one-handed — this is built to run from a phone.

The console on a phone — The console on mobile.

Quick links: mentor.arnao.ai · the app itself.
Anything confusing, wrong, or missing — use the Feedback button inside the app. It lands in the very queue you now manage, so you'll watch your own note arrive. Then reply to Byron's email with what worked, what didn't, and the one thing you'd change first.