Skip to main content
AI & AgentsEarly Access

Approval-first AI store operator

The AI Store Operator is in early access — a copilot that drafts the busywork of running a store and queues it for your approval, never acting on its own. It suggests, you approve, and nothing publishes without your approval.

This is being built

Early Access

The Store Operator is not a shipped feature yet — it is part of our agentic roadmap. The descriptions below are the design we are building toward. Reserve early access and we'll bring you in as it opens.

FaStart's AI Store Operator (early access) is a copilot that drafts catalog enrichment, campaigns, support replies, and late-order flags. Every action runs approval-first: it suggests, you approve, and a full audit trail plus rollback keeps you in control. Nothing publishes without your approval — it does not act on its own.

Operator tasks

The busywork it will draft for you

Each of these is designed to produce a draft you review — never an action it takes alone.

Early Access

Catalog enrichment

Drafts missing descriptions, titles, and attribute fields, prepares SEO-friendly copy, and normalizes product specs. Bulk suggestions, item-by-item approval — nothing publishes without your approval.

Early Access

Campaign drafts

Aims to draft discounts, collections, and announcements from season, stock, and order-history signals. Drafts queue for your review and approval — never an automatic price change.

Early Access

Support reply drafts

Aims to draft replies to order, shipping, and return questions grounded in your store policies. Order context is auto-attached; no message reaches a customer until an operator approves it.

Early Access

Late-order & risk alerts

Plans to surface early warnings for delayed shipments, out-of-stock items, and unusual return patterns. Each flag arrives with a suggested action you choose to approve.

How it works

How the approval-first loop works

The operator's default is "suggest, don't apply". Every task runs through the same four steps, and control always stays with you.

  1. 01

    Suggest

    The operator reads your real catalog and order data and prepares a draft — a campaign, description, reply, or a risk flag.

  2. 02

    Review

    The draft lands in an approval queue with the context for why it was suggested. You can edit, reject, or ask a question.

  3. 03

    Approve

    You give approval, not the operator — one by one or in bulk. Everything approved is logged to the audit trail with who-what-when.

  4. 04

    Apply

    Only after approval does a change go live. A step that goes wrong can be reverted to its prior state in one move via rollback.

By designEarly Access

Nothing publishes without your approval

An operator that runs a real store cannot be a black box. The whole model is built so the AI proposes and a human decides — everything logged, reversible, and inside the boundaries you draw. You set the store's boundaries, not the operator; the AI works inside them.

  • Approval-first: the default is "suggest, don't apply" — nothing goes live without approval.
  • Audit trail: every suggestion, approval, and change is logged with who-what-when.
  • Rollback: an approved change can be reverted to its prior state in one step.
  • Scope & roles: you define which tasks are open to which roles.
Example tasks

Example tasks you could hand the operator

Each one produces a draft — nothing changes until you approve.

Write product descriptions for SEOFill in missing product attributesDraft a weekend discount campaignOrganize products into smart collectionsDraft a reply about a shipping delayFlag products running low on stockSummarize unusual return patternsSuggest a season-aware collection announcement

Frequently asked questions

Does the operator act on my store on its own?

No. The default is "suggest, don't apply". The operator only produces drafts; a campaign, description, or reply never goes live without your approval. You approve, one by one or in bulk.

Is the Store Operator live today?

It is in early access. It is not a shipped feature yet — it is part of our agentic roadmap. The descriptions on this page are the design we are building toward; we onboard waitlisted merchants as it opens.

Can I undo a suggestion after I approve it?

Yes. We aim to let you revert an approved change to its prior state — a campaign or catalog edit that goes wrong can be undone in one step. Every suggestion and approval is also kept in the audit trail.

What does it base its suggestions on?

It is designed to read from the same structured catalog and order data that powers the rest of FaStart, so its drafts reflect your actual products, stock, and policies — not generic templates.

Can I control which tasks are available?

Yes. You define which tasks are open to which roles. You set the store's boundaries, not the operator; the AI works inside them.

Want the operator on your store first?

We're onboarding early-access merchants as the Store Operator takes shape. Join the waitlist and reserve your spot — nothing publishes without your approval.