Shipeasy

Assistant use cases

Three worked chat flows — ramp a feature with an alert on it, generate a measurement plan, and run a read-mode audit of your flags.

Production readyOn this page · 5 min readWorks with · In-dashboard assistant

The assistant is most useful on concrete tasks. Here are three real flows, each showing what you type, what it does, and where the safety rails are.

1. Ramp a feature with an alert watching it

You want the new pricing page behind a flag, ramped slowly, with something paging you if conversion drops while it climbs.

You: "Put the new pricing page behind a flag at 5%, and alert me if conversion falls below 3% over a day."

The assistant scopes the whole thing and surfaces it as confirmation cards you approve — it never just creates resources silently:

It drafts the flag

Name, default-off, and a 5% rollout — pre-filled in an editable card. Change the percentage or add a targeting rule before confirming.

It proposes the metric

A conversion metric over the event you'll fire, created directly once you confirm.

It arms the alert

Comparator, threshold and window on that metric, as one more card. A firing rule files its own item in the queue.

It files the instrumentation

The one thing it can't do from a chat — wiring getFlag(...) and the track call into your code — is filed as a ticket for an agent.

You confirm the cards; the flag is live at 5% with a rule watching it.

write · confirm
Create feature flag
checkout_redesign
Whencountry in [US]
Rollout
10%
Defaultoff

Tweak the title and the highlighted controls here. To change anything else, ask in the chat and the assistant revises the plan.

ConfirmCancel
A write tool surfaces as an editable card — never auto-run. Rules show as chips and the rollout as a bar, exactly like the dashboard; it applies only on Confirm.

2. Generate a measurement plan

You shipped a change and want to know if it worked, but the instrumentation isn't there yet.

You: "Did the new onboarding flow improve activation? Set up whatever we need to measure it."

The assistant produces a measurement plan: it creates the activation metric directly, proposes the flag and the alert rule as cards, and files precise measure_plan tickets for the code it can't write — "fire activated on the first project created", "read step 3's label off getConfig('onboarding_copy')".

suggest_plan
PlanMeasure checkout funnel
1
Create metric checkout_conversion_rate⚡ Instant
count_users(checkout_completed) / count_users(checkout_started)
2
Create experiment new-checkout-flow⚡ Instant
control 50% · variant 50% · goal ↑ checkout_conversion_rate
3
Instrument checkout_completed⚒ Ship run
Emit on the order-confirmation screen.
1 code edit → filed to your ship crew · 2 instant
Approve planDismiss
A plan separates Instant resources (created on approve) from Ship-code work (filed as a ticket). One approvable unit for a whole measurement.
Why a ticket and not an edit

The in-dashboard assistant has no repo access. A measurement plan is its way of being complete: it builds the parts it can and hands off an unambiguous spec for the parts it can't, so a developer or ops:work agent can finish the wiring.

3. Read-mode audit of your flags

Sometimes you just want answers, not changes. In read mode the assistant executes read tools directly — no confirmation cards, because nothing is mutated.

You: "Which feature flags are still at 100% but haven't been touched in 90 days? Any I can retire?"

It queries your flags, configs and kill switches and answers in the chat — listing flags stuck at 100% for months, configs no targeting rule references, and kill switches nobody has exercised. It cites what it found; you decide what to clean up.

read · auto
Which feature flags are at 100% but haven’t been touched in 90 days?
Reading your project
Checking experiment status
Searching the docs
Three flags are pinned at 100% and untouched for 90+ days — safe to retire:
  • legacy_checkout — fully rolled out, no rule references it
  • dark_mode — at 100% since March, experiment finished
  • beta_banner — superseded by nux_banner
Sources: Retiring flags · Flag lifecycle
Read tools run on their own and stream as a ✓ trace; the answer is grounded in your project and cites its docs. Nothing changes.

Prop

Type

This read/write split is the assistant's core safety model: reads are free, writes are always confirmed.

Was this page helpful?
Updated August 9, 2026✎ Edit this page

On this page