BotSupport vs Maisie AI: which ecommerce chat assistant actually fits your store?

What is Maisie AI, and can you still buy it today?

Maisie is best understood as a marketing-led chatbot builder rather than a helpdesk. Its central idea was that a shop assistant conversation converts better than a promotional banner, so the product leaned into engagement mechanics: conversations triggered by time on page, exit intent, cart value or shipping thresholds, quizzes that narrow a shopper down to a product, and opt-in prompts that push captured emails and phone numbers into Klaviyo. Support automation sat alongside that, with AI-generated answers over the content the merchant supplied, real-time order status, and escalation to a human via live chat, Messenger or email. That is a coherent product with a clear point of view, and it is worth saying so before we get to where we differ.

The build model was do-it-yourself. Independent listings describe a library of prebuilt Shopify conversation templates plus a custom drag-and-drop flow builder, so a merchant would start from a template, edit the branches, write the answers and publish. That is a legitimate approach and it suits people who want to own the wording. It also means the quality of the assistant is a direct function of how many hours you put into it, and how diligently you revisit it when products, prices and policies change. Nothing about that is a criticism of the tool. It is simply a different distribution of work, and it is the distribution that decides most purchases in this category.

On traction, we are going to be deliberately vague, because the honest position is that we do not know. Third-party install trackers, which detect only the subset of stores they can fingerprint and systematically undercount, showed a small and declining Shopify footprint through mid-2026. Vendor-confirmed figures are not public. Its Shopify App Store rating was high, but from a small number of reviews, so read it as encouraging rather than statistically meaningful. We would ask you to apply exactly the same scepticism to us: BotSupport is newer still and has no meaningful install base at all.

The availability point matters more than any feature comparison, so we want to be precise about what we observed rather than what we infer. When we checked on 6 August 2026, we requested maisieai.com and www.maisieai.com directly and both returned a 404, as did the pricing, about and integrations pages. The BigCommerce app page returned a 404. The Shopify App Store page still loads, but carries the notice: "This app is not currently available on the Shopify App Store. If you have support questions, contact Maisie Technologies directly." A third-party tracker also showed the Shopify listing as no longer live when we checked, and trackers lag and misfire, so treat that as a prompt to check rather than a fact. We do not know why. Sites go down for all sorts of ordinary reasons, from a billing hiccup to a replatform, so we are not going to guess. If you are evaluating Maisie, check maisieai.com yourself before you spend time on it. If you are an existing customer, use the support details on the Shopify App Store listing.

• Category: self-serve ecommerce chatbot and live chat builder, with a quiz and lead-capture emphasis
• Native apps: Shopify and BigCommerce, as documented; we found no native WooCommerce integration
• Pricing basis: monthly unique website visitors, with a 14 day free trial, per independent listings
• Publisher: Maisie Technologies, an Australian company, according to its Shopify App Store listing
• Last checked 6 August 2026: the vendor website did not resolve for us and the Shopify listing showed the app as not currently available

Who actually builds and maintains the assistant?

This is the axis that decides most of these purchases, and it is worth being blunt about it. A flow builder is not the hard part of a chatbot. The hard part is the content: knowing that your delivery cut-off is 2pm not 3pm, that the size guide for one brand runs small, that returns on sale items work differently, and that a customer asking "will this fit a 12 month old" is really asking about two product attributes at once. Somebody has to sit down and encode that. Maisie's model is that the somebody is you. BotSupport's model is that the somebody is us. Neither is inherently better. They suit different constraints, and the constraint that usually decides it is whether your scarce resource is money or time.

The second half of the difference is upkeep, and it is the part buyers consistently underestimate. A self-built flow can degrade quietly, unless something in your process catches it. Products get discontinued, a policy page gets rewritten, a courier changes, and the assistant may carry on repeating last quarter's answer. Often nothing alerts you, because from the tool's point of view nothing has broken. With a managed service, keeping the assistant aligned with the store is part of what you are paying for rather than a recurring task on someone's list that never quite gets prioritised. If you already have a disciplined content review process, that advantage shrinks considerably, and you should discount it accordingly.

We should be fair about the trade-off in the other direction. A DIY builder gives you immediate control. If you want to change a line of copy at eleven at night, you log in and change it. With a managed service you ask, and you wait for a person, even if that wait is short. For some teams that dependency is genuinely worse than doing the work themselves, and we would rather say so than pretend the managed model is free of downsides. There is also a question of institutional knowledge. If your team builds the assistant, your team understands it. If we build it, the understanding sits with us, and you are relying on us to keep it. That is a real consideration for anyone who has been burned by a supplier before.

Does Maisie work with WooCommerce, or is it Shopify-first?

Platform limits in this category are rarely advertised clearly, so it is worth naming the specific pattern. Chat widgets are easy to put on any site, because a script tag is a script tag. What is not portable is the commerce layer. Pulling a customer's order status, reading live stock, or recommending a product with a current price all require an authenticated connection to the store, and that connection is built per platform. When a vendor lists a Shopify app and a BigCommerce app and nothing else, the reasonable reading is that a WooCommerce shop is likely to get a chat bubble and static FAQ answers, but we found no supported route to order lookups or live catalogue recommendations on WooCommerce. There may be a webhook or custom integration path we did not find, and if you are a WooCommerce merchant that is the first question to ask any vendor in this space.

For UK independent retailers this matters disproportionately, because WooCommerce remains very common among smaller and mid-sized British stores that started on WordPress. Being told to migrate platform in order to use a chatbot is not a serious answer for most of those businesses. BotSupport was built with that in mind: it runs on WooCommerce and Shopify equally, and on sites that are on neither, because the grounding work is done by a person against your actual catalogue and policy pages rather than assumed from a single platform API. The honest limit on our side is that some setups resist this. Headless builds, storefronts with strict content security policies, and platforms with locked templates can all get in the way. If we cannot ground the assistant properly on your setup, we will tell you before you pay.

Where Maisie was clearly the better fit on platform terms was BigCommerce. It had a genuine native BigCommerce app at a time when many ecommerce chat tools ignore the platform entirely. If your store is on BigCommerce, that was a real and specific advantage, and it is the sort of thing a comparison article should say plainly rather than bury. It is also worth noting that at the time of writing that BigCommerce listing page was not loading for us, so if BigCommerce support is your deciding factor, confirm the current position with the vendor before you plan around it.

What does it mean for a chat assistant to be grounded in your live catalogue?

The practical test is simple, and you should run it on any tool you are trialling, including ours. Ask about a specific variant of a specific product, then ask a follow-up that combines two attributes, then ask something your policy page covers in a slightly awkward way. For example: "do you do the navy one in a 14", then "is it in stock", then "if it does not fit can I exchange rather than refund". A well-grounded assistant should handle all three. An assistant working from pasted text will usually handle whichever of those its author anticipated, and will be as current as the last time that text was updated. That is the risk to test for, and it is testable in ten minutes.

This is also where the DIY versus managed split shows up technically rather than just commercially. Encoding a catalogue properly is fiddly. It involves deciding what the assistant should do when a product is out of stock, how to phrase price answers when there are sale prices running, how to handle pre-orders and backorders, and when to stop guessing and hand over to a human. A self-serve builder exposes those decisions to the merchant as configuration, which is honest but time-consuming. A managed build makes them on your behalf, then shows you the result and adjusts it. If you have strong opinions about all of those decisions, the self-serve route may genuinely suit you better.

We should also be clear about the limits of our own claim. No assistant is right every time. Large language models make mistakes, catalogues contain contradictions, and policy pages are often ambiguous in ways nobody notices until a customer asks. Grounding reduces a specific class of error, the stale answer, and it makes handover behaviour easier to get right. It does not eliminate error. Test ours with the twenty-question method further down before you commit, and judge it on what you see rather than on what we say here.

Being fair to Maisie on this point: nothing about its architecture prevented good grounding, and the reviews we saw were positive, though there were few of them. The constraint we would point to is effort rather than capability. A merchant with the time to encode their catalogue carefully into a flow builder can produce something very good. The question is whether that merchant is you, and whether those hours are better spent elsewhere.

How do the pricing models compare, and what does the staff time really cost?

Visitor-metered pricing has one obvious characteristic, and it is the pricing model rather than the specific numbers that matters for your decision. Your bill rises with traffic whether or not that traffic buys anything. A seasonal spike, a press mention or a paid campaign that underperforms all push you up a tier. The flip side is real and in Maisie's favour: the entry point was genuinely cheap, which makes it a low-risk way for a small store to find out whether a chatbot helps at all. Listings also described generous conversation and bot allowances even on lower tiers, which is unusual and worth crediting. As with most tiered tools, check the current plan list before assuming any particular feature is on the plan you are considering.

The cost buyers routinely forget is not on either price list. It is the hours. Building a decent set of flows, writing the answer content, testing it, and then maintaining it is real work by someone who understands the products. We are not going to invent an average figure for that, because it varies enormously between a fifty-product boutique and a five-thousand-product catalogue. Do the arithmetic for your own organisation instead: take the loaded hourly cost of the person who would build it, estimate the hours for the initial build, add a monthly maintenance allowance, and add that to the licence fee. For most small teams the answer is that the software is the cheap part and the labour is the expensive part. That is the entire commercial argument for a managed service, and if the arithmetic comes out the other way for you, the managed service is the wrong purchase.

BotSupport's flat GBP 79 per month at Early Access includes that labour. It does not scale with traffic, so a good month does not cost you more, and a quiet month does not cost you less. It is also, honestly, more than a low-traffic self-serve tier. If you have a thousand visitors a month and plenty of spare time, a cheap tool you build yourself is the cheaper answer, and we would rather you reach that conclusion here than after a trial. Early Access pricing is also, by definition, introductory. We would expect it to change as the product matures, and anyone signing up now should ask us what happens to their rate when it does.

How long does it take to get a working assistant live?

Time to a widget and time to a working assistant are two different metrics, and self-serve tools are usually marketed on the first. Installing an app and switching on a template is fast, and vendors are right to say so. The gap opens up afterwards, in the unglamorous stretch where you discover the assistant cannot answer half of what customers actually ask, and you go back into the builder repeatedly to close the gaps. That stretch is where most self-serve chatbot projects quietly stall, not because the tool failed but because the person who was going to write the content got busy.

The managed model compresses that stretch by moving it off your calendar. You supply access and the policy detail, a person builds and tests, and what comes back is intended to be usable on day one rather than a starting point. It is a different shape of commitment: less of your time overall, but a hard dependency on someone else's turnaround for changes. If your business changes its offers weekly and you need same-hour edits, ask us specifically how we handle that before you sign up, because it is the scenario where the managed model is least comfortable.

Whichever you choose, judge the result on the same evidence. Go live, then read the first hundred real conversations. That is the only measure that matters, and it is available to you within a fortnight either way. Count how many questions were answered correctly, how many were answered wrongly, and how many were handed to a human. The third number is not a failure. A confident wrong answer costs far more than an honest escalation, and any assistant that never escalates is either extraordinary or overconfident.

Who should honestly choose Maisie, or a Maisie-style tool, instead?

The genuinely strong parts of Maisie were not things BotSupport does. Product recommendation quizzes are a marketing mechanic, not a support feature, and they are effective for stores where choosing between options is the real friction. Native Klaviyo sync for captured emails and phone numbers is a proper integration that saves real work for a marketing team that already lives in Klaviyo. Facebook Messenger as a channel matters to some audiences more than any website widget does. Behavioural triggers on exit intent, cart value and time on page are conversion tooling, and they work. If your brief is "grow the list and lift conversion", that is a marketing chatbot brief, and BotSupport is not primarily a marketing chatbot.

BotSupport's brief is narrower and different: answer customer questions accurately from the live catalogue, around the clock, without your team building or maintaining it. Those are complementary jobs rather than competing ones, and a store with money and time might reasonably want both. We would rather you understood that distinction than bought us for a job we do not do, because the refund conversation is worse for everyone than the honest conversation up front.

On alternatives, and we say this because a comparison that only ever points at itself is not useful: if you specifically want the quiz mechanic, look at quiz-led tools such as Octane AI. If you want a self-serve widget with live chat, self-serve options such as Tidio are a common starting point. If your problem is genuinely a ticket volume problem across email, chat and social, you are looking at a helpdesk such as Gorgias rather than a widget. Check their current plans and free-tier terms yourself, as those change frequently and anything we print here will date. None of those are BotSupport, and pretending otherwise would not help you.

How do I trial both, or move off Maisie if I need to?

A structured trial is worth an afternoon of preparation. Pull the last hundred customer emails or chat transcripts and group them: order status, sizing and fit, stock and availability, delivery and cut-offs, returns and exchanges, product comparison, and everything else. Pick your top twenty by frequency, then deliberately add two that are awkward, the ones your own staff have to think about. Score each answer as correct, partly correct or wrong, and separately note whether the assistant handed over to a human when it should have. The handover behaviour matters as much as the answers. An assistant that says "I am not sure, let me get someone" is doing its job; one that invents a delivery date is not.

If you are an existing Maisie customer and the tool is not currently reachable for you, the practical steps are unglamorous but important. Copy your conversation flows and answer text out into a document while you can still reach the admin, because that content is the asset and it is portable even if the tool is not. Check what is still injecting a script into your storefront theme and remove anything that is no longer loading, so customers are not clicking a chat bubble that never opens. Check your Klaviyo lists to confirm capture has not silently stopped, because a broken opt-in is invisible from the front end. Then choose a replacement on its merits rather than on urgency, which is the hardest part of any forced migration.

Moving that content into BotSupport is straightforward because the build is done for you. You hand over the flow copy and policy detail you exported, we build and test the assistant against the live catalogue, and it comes back ready for you to review. There is no flow rebuilding on your side and no migration tooling to learn, which is the practical benefit of the managed model at exactly the moment a self-serve tool has become inconvenient. We would still expect you to review what we produce and tell us what is wrong, because you know your customers and we do not, yet.

Where is BotSupport the weaker choice?

We would rather list this plainly than have you find out during a trial. Early Access means what it says. BotSupport is a young product from a UK team, and if your procurement process requires reference customers at your scale, we do not have a list to give you. Being an early customer has an upside, which is that you get an unusual amount of attention and your feedback actually changes the product, and a downside, which is that you are carrying risk a later customer will not have to carry. Price that in honestly. If you cannot afford for this to go wrong, wait six months and ask us again.

The managed model also concentrates risk in the provider. You are buying a service relationship, and that is only as good as the responsiveness behind it. If we are slow, you are slow, and there is no self-serve escape hatch to change a line of copy yourself. We think the trade is worth it for most small retail teams, because their genuine scarcity is time rather than software budget, but it is a trade and not a free win.

What we are confident about is the specific commitment we make: a person builds the assistant on your real products, prices and policies, tests it, and hands it over working within one working day, on WooCommerce, Shopify or most other websites, and we will tell you before you pay if your setup is one we cannot ground properly. If that is the problem you have, the comparison is favourable. If your problem is list growth, quizzes or multi-channel ticket volume, it is not, and we have named better places to look above.

Sources

1. Maisie AI Chatbots & Live Chat, Shopify App Store listing — Retrieved and re-verified 6 August 2026 (HTTP 200). Page loads but states: "This app is not currently available on the Shopify App Store. If you have support questions, contact Maisie Technologies directly." Also names Maisie Technologies as the developer and support contact.
2. Digismoothie: Maisie app for Shopify, review, pros and cons — Independent review, retrieved 6 August 2026 (HTTP 200). Records visitor-metered pricing from around USD 15 per month rising with traffic bands and custom pricing above high volumes, a 14 day free trial, no free plan, a library of chatbot templates plus a drag-and-drop builder, Messenger and Klaviyo integrations, and all features included with unlimited bots and conversations. Flags rising cost with traffic as a drawback.
3. Store Leads: Maisie AI Chatbots & Live Chat Shopify app report — Third-party install tracking, retrieved 6 August 2026 (HTTP 200). Used only for the directional point that the detected Shopify footprint was small and declining, and that the tracker showed the listing as no longer live. Trackers detect a subset of stores, lag and misfire; figures are not vendor-confirmed.
4. PickYourApp listing: Maisie AI Chatbots & Live Chat — Aggregator snapshot of the Shopify listing, retrieved 6 August 2026 (HTTP 200). Used for feature description only: prebuilt Shopify conversation templates, visitor-based plan structure, Klaviyo integration. Aggregator data can be out of date.
5. Maisie AI company page, LinkedIn — Vendor self-reported profile, retrieved 6 August 2026 (HTTP 200). Used only to corroborate that Maisie Technologies is an Australian company that has been in the conversational commerce category since around 2018.
6. maisieai.com (vendor website) — Requested directly on 6 August 2026. Both maisieai.com and www.maisieai.com returned HTTP 404, as did /pricing, /about and /integrations/bigcommerce-chatbot. We could not confirm why. Check again before relying on this.
7. Maisie AI Chatbots + Live Chat + Quizzes, BigCommerce app marketplace — Previously the native BigCommerce app listing. Returned HTTP 404 when requested on 6 August 2026.

Details on Maisie AI reflect its public documentation and independent reviews as of August 2026 and change frequently. This comparison is published by BotSupport, so read it as a vendor comparison and check the alternative for yourself. If something here is out of date, tell us and we will correct it.