Meta launched Muse — its AI assistant for small businesses — on 29 September 2026, and the questions landing in our inbox have been piling up ever since. Is it any good? What does it actually do that ChatGPT or Claude doesn’t? Can UK operators use it yet? Should you connect your Shopify store? This is the practical guide we wish we’d had in the first fortnight, written for owner-operators and small teams rather than enterprise buyers.
One thing to clear up before we start, because the naming is genuinely confusing: Muse here means Meta’s Muse. There is a Microsoft research model called Muse and a handful of older Google projects that used the name. This article is not about those. This is Meta’s Muse for Small Business, the one that sits inside the Meta Business Suite and connects to QuickBooks, Shopify, Slack and Meta Ads Manager.
Who can actually use Meta Muse right now
As of 29 September 2026, Muse is available to small and medium businesses in the United States and Canada. You need a Meta Business account in good standing and you sign up through the Muse for Small Business waitlist inside Meta Business Suite. Access has been rolling out in waves — new accounts typically get in within a week or two.
A UK and EU rollout has been confirmed as planned, but Meta has not published a date. In private briefings they have pointed to “2027” without committing to a quarter. Realistically, UK readers should plan for availability in the first half of 2027 and treat anything earlier as a bonus. We’ll come back at the end to what to do in the meantime, because you do not need to sit on your hands.
Eligibility inside the US and Canada is loose. Sole traders with an active Meta ad account can sign up alongside businesses of 100 people. The platform is clearly aimed at the long tail of SMBs that Meta already monetises through advertising rather than at enterprise accounts — which, for once, is a good thing. It means the defaults and the UI are built for people who do not have an IT team.
Getting access: the actual steps
The signup is straightforward, but worth walking through because the connector permissions screens matter and most writeups skip them.
- Step 1. Sign in to Meta Business Suite with the account that owns your Facebook Page or Instagram business profile. Muse lives under the “AI tools” section in the left-hand nav — it appears once your account is eligible.
- Step 2. Click “Join Muse for Small Business.” You’ll be asked a handful of questions about your business type, headcount band and primary goal. Meta uses these to personalise the onboarding prompts; the answers are not binding.
- Step 3. Accept the Muse-specific terms. These are separate from your existing Meta Business terms. Read the section on data use — it’s short and worth understanding before you connect anything.
- Step 4. You’ll land in the Muse workspace with a blank chat and a “Connect your tools” panel on the right. Do not connect everything at once. Start with one tool, test it, then add the next.
Access is free at the time of writing. Meta has indicated that Muse will eventually move to a paid tier for heavier usage, with the free tier capped at a monthly message limit. Pricing has not been announced.
Connecting your tools: what each connector actually pulls
This is the part most coverage glosses over. “Connects to QuickBooks” means very different things depending on which scopes you grant. Here is what we’ve seen each connector request as of November 2026, and what it does with that data.
QuickBooks connector
Pulls your chart of accounts, invoices (issued and received), expenses, bank feed transactions and the standard P&L and balance sheet reports. It does not have write access — Muse cannot create or edit QuickBooks records. You use it to ask questions like “which five customers are more than 30 days overdue” or “what did we spend on Meta Ads last quarter versus the one before.” It reads live data; you don’t have to export anything.
Shopify connector
Pulls product catalogue (titles, descriptions, variants, prices, images), orders, customers (with PII masked by default — you can toggle), and store analytics. Again, read-only at launch. The headline use is reconciling orders against QuickBooks, pulling “best sellers this week” style summaries, and feeding product details into ad copy prompts without you having to paste anything.
Slack connector
You pick the channels. Muse can read message history in the channels you choose and can be @-mentioned to answer questions in them. It does not read DMs. Common pattern: connect your #orders, #support and #ops channels, then ask Muse to summarise what happened overnight or extract action items from a long thread.
Meta Ads Manager connector
This is the connector where Muse has a genuine advantage over general-purpose chatbots. It pulls campaign structure, creative, spend, impressions, clicks, CPA and ROAS across Facebook and Instagram ads. You can ask “which of my current ads is underperforming and why” and get an answer grounded in your actual numbers rather than hallucinated averages.
WhatsApp Business and Instagram
These are available if you already use the Meta Business surfaces, but with narrower scopes — mainly for drafting replies and summarising inbox traffic, not for autonomous messaging. We’d treat these as nice-to-have rather than reasons to adopt Muse.
Practical use cases that earn their keep
The following are the four jobs we’ve seen pay back the setup time fastest. If you only get Muse to do these, it’s worth the hour it takes to set up.
1. Reconcile sales across Shopify and QuickBooks
The classic end-of-week finance chore. Instead of exporting two CSVs and joining them in a spreadsheet, you ask Muse to compare Shopify orders for the week against QuickBooks sales receipts and flag anything that doesn’t line up. It handles timezone quirks, refunds and partial captures surprisingly well. You still want a human to approve the final numbers, but the first pass takes seconds rather than an hour.
2. Draft Instagram ad copy using real product data
Because Muse can see both your Shopify catalogue and your Ads Manager performance history, it can draft copy that references actual product names, prices and benefits, and lean on hooks that have worked in your previous ads. This is the single most time-saving writing task we’ve seen for e-commerce operators.
3. Summarise Slack conversations
Especially useful on Monday mornings or after a day off. @-mention Muse in a channel and ask for a summary of the last 48 hours with decisions, open questions and anything flagged as urgent. The quality is close to a decent human note-taker, with the caveat that it will miss subtext — so treat it as a first pass, not the final word.
4. Weekly revenue and ad spend summary
Pull a single weekly briefing that combines Shopify revenue, QuickBooks costs and Meta Ads spend into one short report. We have clients who now start Monday with a five-line Muse summary instead of three open tabs and a nagging feeling they’re missing something.
Prompts to copy and paste
These are ten prompts we’ve tested across real SMB accounts. They assume you’ve connected the relevant tools. Change the square brackets to fit your business and you’re off.
Give me a weekly briefing covering [last week].
Include: total Shopify revenue, number of orders, average order value, Meta Ads spend, blended ROAS, and the single biggest change versus the previous week.
Keep it to five lines. Flag anything I should look at today.Compare Shopify orders for [date range] against QuickBooks sales receipts for the same period.
List any orders in Shopify that are missing from QuickBooks, and any receipts in QuickBooks with no matching Shopify order.
Ignore refunds under £10 and include notes on anything in a currency other than GBP.Draft three Instagram ad captions for [product name] aimed at [audience].
Use the actual product benefits from Shopify, not generic claims.
Match the tone of my three best-performing ads from the last 90 days.
Each caption: under 120 characters, one emoji maximum, clear call to action.From QuickBooks, list customers with invoices more than 30 days overdue.
Sort by amount owed, highest first.
For each: customer name, total outstanding, oldest invoice date, and a suggested next step based on how overdue they are.Summarise activity in #orders, #support and #ops over the last 24 hours.
Group into: decisions made, open questions, items flagged urgent, customer issues needing follow-up.
Name the person responsible for each open item if it’s clear from the thread.Look at my active Meta Ads campaigns.
Identify the three worst-performing ads by CPA over the last 14 days.
For each: name, CPA, ROAS, what’s different about it compared to my best performers, and one specific thing to test.What are my top ten Shopify products by revenue for [date range]?
Include units sold, revenue, gross margin if available, and whether each is trending up or down versus the previous period of the same length.From QuickBooks, show me expense totals by category for [month] versus [previous month].
Highlight any category that moved by more than 20% in either direction.
Suggest two questions I should ask myself about the biggest increase.Draft a one-page monthly update for [month] in the tone of a founder writing to a small advisory board.
Cover: revenue, gross margin, cash position signals from QuickBooks, Meta Ads efficiency, one win, one challenge, one ask.
Honest tone. No marketing fluff.Read the pinned thread in #ops.
Extract every commitment made, who made it, and the implied deadline.
Format as a checklist I can paste back into Slack.What NOT to use Meta Muse for yet
This is where most launch coverage is a bit breathless. Muse is useful, but it is not a replacement for human judgement in the places that matter most. A few concrete lines we’d draw:
Customer-facing conversations without a human in the loop. Muse can draft a reply to a support ticket. It should not send one unreviewed. Meta positions Muse explicitly as a draft generator and task assistant, not as an autonomous customer service agent, and that framing is correct. Treat its output as a first draft that a human signs off.
Autonomous actions that spend money or ship orders. The write-access connectors are not live yet, and even when they arrive, we’d be slow to grant them. If a prompt mistake could issue a refund, change an ad budget or ship a wrong SKU, keep the human in the loop. The time saved by removing the approval click is never worth the risk of a hallucinated £5,000 refund.
Confidential data that isn’t already in a connected system. Don’t paste employment contracts, legal matters, M&A details or anything covered by a strict DPA. The connector data is covered by Meta’s small business terms; pasted text is more ambiguous. If in doubt, redact.
Anything where you need an audit trail you’d defend to HMRC. Use Muse to draft, summarise and investigate. Use your accounting system of record for the actual numbers you file on.
UK readers: what to do while you wait
If you’re reading this from the UK, the honest position is that Muse isn’t available to you yet and nobody at Meta will commit to a date. That’s frustrating, but it is not a reason to wait. Almost every workflow we’ve described above can be done today with a general-purpose AI tool, some manual exports and a bit of setup.
For the writing and summarising jobs — ad copy, weekly briefings, Slack summaries, board updates — Claude or ChatGPT will do at least as well as Muse. You just have to bring the data in yourself. A Shopify CSV export and a QuickBooks P&L pasted into a Claude Project gives you 80% of what Muse does, and you can set it up this afternoon.
For the live-data jobs — the reconciliation and the ad diagnosis — the honest answer is that general-purpose chatbots need you to export the data first. For most SMBs a weekly export habit is fine. For teams that need something closer to Muse’s live-data experience, an MCP server or a Zapier bridge into Claude gets you there, with more setup effort than Muse but roughly the same end result.
Our recommendation for UK operators: build the prompts and the habits now with Claude or ChatGPT, learn what you actually want an AI assistant to do for your business, and switch some of the workflows to Muse when it lands. The muscle memory transfers. The prompts above will all work, with minor tweaks, in Claude today.
If you want help setting up those workflows the way we’d set them up for a client, that’s what we do — see our training and implementation options, or book a 30-minute call and we’ll walk through your specific stack.
The bottom line on Meta Muse
Muse is a credible first move by Meta into assistant-style AI for the SMB market. The native connectors to Shopify, QuickBooks and Ads Manager are the genuine differentiator — not the underlying model quality, which is comparable to other 2026 assistants. For US and Canadian operators already running on that stack, it’s worth the hour to set up and the ten prompts above will give you an immediate feel for whether it belongs in your week.
For UK operators, it’s a signal of where the market is going more than a tool you can use today. Build the habits now with what you’ve got, and you’ll be ready when the UK rollout lands.