All posts
Ernest Team10 min read

Crisp Alternative for Shopify and WooCommerce: Inbox With a Chatbot vs Store-Connected Agent

Crisp is a well-built multichannel inbox for small SaaS teams. A Crisp alternative for stores that need live order lookup, catalog answers, and returns.

A customer messages at 11pm asking where order 1043 is. In Crisp, that conversation lands in a genuinely nice inbox that already knows who they are, because the Shopify integration has synced their data into the sidebar. What happens next depends entirely on whether someone is awake to read it.

That gap, between having the customer's record visible in the inbox and having something that can go look up the order and answer, is what sends store owners searching for a Crisp alternative. It usually isn't dissatisfaction with Crisp as software. Crisp is well designed and priced kindly at the entry point. The searches come from a mismatch in shape: Crisp is a shared inbox that a chatbot plugs into, and an ecommerce store increasingly wants an agent that is plugged into the store.

What Crisp is good at

Crisp deserves a fair hearing before any of the criticism, because most people evaluating it are already half-sold for reasons that hold up.

The free plan is real. Per Crisp's pricing page, as of publication, Free costs $0 forever and includes 2 seats, the website chat widget, a shared inbox, mobile apps, chat SDKs, contact forms, push notifications, and the ecommerce integrations. Two seats and a shared inbox at $0 covers a two-person business completely, and the plan card lists the Shopify, Adobe Commerce, WooCommerce and Prestashop connections as included at that tier rather than held back for a paid one.

The pricing philosophy is the other draw. Crisp bills a flat rate per workspace and states plainly on the page that you pay a single consistent price regardless of how many conversations you have in a month. For a store with seasonal spikes, that removes the specific anxiety of watching a conversation counter during Black Friday. Crisp is explicit about this being a deliberate stance against consumption billing.

Developers like it for a concrete reason: chat SDKs sit on the Free plan card and developer APIs are on the feature list, and one of the testimonials Crisp publishes describes a customer using the API enough that Crisp became a mini back-office for them. That is a specific kind of praise, and it is the kind that small SaaS teams respond to. Crisp also lists EU hosting, TLS encryption, a data processing agreement and data replication in its security section, which matters if your buyers ask GDPR questions.

There is also more AI in the product than "chatbot" suggests. Crisp's own feature list includes an AI Agent described as handling conversations end to end, AI knowledge training on internal data, an internal copilot for agents, multilingual AI, auto-tagging, and conversation summaries. Crisp says 10,000 companies use it. Every plan comes with a 14-day trial of all features, no card required.

Crisp's pricing, in Crisp's own numbers

Four published tiers, all billed per workspace per month, as of publication:

PlanPrice/moSeats includedAI creditsCustomer profiles
Free$02$0 (none)100
Mini$454$5, around 90 automated conversations5,000
Essentials$9510$25, around 450 automated conversations50,000
Plus$29520+$75, around 1,350 automated conversations200,000

Two things in that table matter more than the headline prices.

First, conversations are unlimited but automated conversations are not. AI is metered in dollars of credit, and Crisp's own page pairs each credit allowance with an approximate conversation count. Divide their two numbers and $5 of credit against roughly 90 automated conversations lands near five or six cents per conversation, though that is arithmetic on Crisp's figures rather than a rate Crisp publishes. The flat-fee promise covers human conversation volume. AI volume is consumption billing living inside a flat-fee plan.

Second, the AI budget scales with the plan, not with your traffic. A store on Mini at $45 gets AI coverage for about 90 conversations. If it handles 600 support conversations a month and wants most of them answered automatically, the published path to more AI credit is a higher tier, which also buys seats and customer profiles it may not need.

Crisp lists additional agents at $10 per month on the Plus plan. Sub-inboxes run 0 on Free and Mini, 2 on Essentials, 5 on Plus. Discounts published on the page: 50% for nonprofits and students, and 20% off for three additional workspaces on Plus and Essentials. VAT is not included in the listed prices, though EU businesses can zero it out with a VAT number and non-EU customers aren't charged it.

What Crisp does not publish there

Worth stating plainly rather than guessing at:

  • The price of extra AI credits. If you burn through $25 of credit on Essentials in week two, the pricing page doesn't say what the next block costs.
  • Additional seat pricing below Plus. Only the Plus column carries the $10 figure. Free, Mini and Essentials show nothing.
  • The annual discount. Crisp confirms yearly billing exists and is toggled inside the app after subscribing, but the page gives no percentage.
  • Enterprise pricing. Listed as "unique pricing" with a demo request.
  • The cost of extra channels. Crisp's FAQ says a workspace holds one of each channel (Instagram, Messenger, WhatsApp Business, Viber, Line, Telegram) and that you connect more "by subscribing to higher limits". No figure is attached to that.

Much of Crisp's pricing page renders client-side and a plain fetch doesn't capture the per-feature checkmarks, so this article deliberately avoids claiming which tier a given feature starts at unless Crisp's plan card names it directly. Check the live page before you budget against any of this.

Where the fit breaks for a store

Crisp's plan cards do tell you where things sit. The omnichannel AI chatbot and the workflow automation builder appear under Essentials at $95. Task automations that connect the AI agent to third-party tools appear under Plus at $295, alongside the ticketing system and 100+ integrations.

Read that as a sequence and the shape of the work becomes clear. Getting a Crisp AI agent to answer "where is order 1043" with a real answer means building it: a workflow in the visual builder, wired to an automation that calls out to Shopify, on the tier where task automations are included. Crisp gives you good tools to build that. It is still a build.

Crisp describes its ecommerce integrations as syncing conversations and order data with your inbox so agents get context. That is a different job from an agent holding live catalog and order access as tools it can call mid-sentence. Crisp's page doesn't claim out-of-the-box order lookup by the AI agent, and this article won't claim it can't be built. What it will say is that the pricing page positions the integration as context for the humans and the builder as the path to anything more.

The same holds for the three questions that dominate ecommerce chat volume:

"Is this in stock in a medium?" needs the live catalog at answer time, not a synced customer profile.

"Cancel my order, I ordered the wrong size." needs the agent to take an action against the store, with a rule about whether a human approves it.

"I want to return this." needs the return policy and the order together, plus a decision about who is allowed to start the return.

If those are most of your queue, you're buying an inbox and then building the store-connected part yourself. That is a fine trade for a team with an engineer and a preference for control. It is a poor trade for a three-person store that wanted the answers, not the toolkit.

Who should stay on Crisp

Stay if you're a small SaaS or service business rather than a store, and your support is mostly humans in a shared inbox with AI as a helper. Stay if channel breadth is the requirement, because Crisp's list runs to WhatsApp Business, Instagram, Messenger, SMS, Telegram, Viber, Line, X and phone, and no ecommerce-first AI tool matches that. Stay if you want a help center, a status page, a customer portal and a CRM from one vendor. Stay if two seats at $0 is exactly the deal you need this quarter.

Where Ernest fits

Ernest starts from the store rather than from the inbox. It installs on Shopify or WooCommerce, imports policies and catalog automatically on connect, and scrapes the rest of the site for FAQs and product detail. Order lookup and product search are tools the agent calls during the conversation, not workflows you assemble. Cancellations and returns can run fully automatic or require one-click approval, per action, merchant's choice. Refunds are always approval-gated, with no setting to change that.

Before it answers a real customer, you can replay your own past tickets against it in simulation and read how it would have handled them. Once it's live, conversation intelligence surfaces the questions that keep coming back with no good answer. It replies 24/7 in the customer's language.

Pricing is volume-based and one number: Free at $0 for 100 conversations that never expire (a free plan, not a trial), then $49/month for 500, $149/month for 2,000, $299/month for 10,000. AI is included at every tier. There are no per-seat fees and no separate AI credit balance, so the metered-AI-inside-a-flat-fee split doesn't exist here. The trade is the mirror image: your conversation volume is capped and your AI usage isn't, where Crisp caps AI usage and doesn't cap conversations.

What Ernest does not do, stated plainly because a comparison without this is worthless:

  • No phone, SMS, WhatsApp, Instagram, or Messenger. Storefront widget and email, with escalation to a human by email. Crisp's channel list is far longer and that is not close.
  • No shared-inbox workflows. No sub-inboxes, no round-robin routing, no assignment rules, no ticketing portal. If your support model is several people dividing a queue, Crisp is built for that and Ernest is not.
  • No CRM. No customer profiles, segmentation, lifecycle tracking or timeline views. No helpdesk, ERP, WMS or returns-portal integrations either.
  • No white labelling. Crisp sells watermark removal on Plus; Ernest doesn't offer it at all.
  • No autonomous refunds. If you want money moving without a click, Ernest will disappoint you on purpose.

Two questions that settle it

Is your support queue mostly store questions? Count last month. If order status, stock, sizing, cancellations and returns are the bulk of it, you want an agent with store access, and the version of that you build inside a general inbox will cost you engineering time on top of the subscription. If your queue is billing questions, bug reports and onboarding help, Crisp's shape is the right one and switching would be a downgrade.

Which cap hurts less? Crisp gives unlimited conversations and meters the AI. Ernest gives unlimited AI within a conversation cap. Work out your monthly conversation count and what share you want answered without a human, and one of those two models will obviously fit your numbers better than the other.

If you're comparing more than these two, our roundup of customer service chatbots covers the wider field, and the Tidio alternative comparison walks through another tool that splits its chat plan from its AI billing. For the specific mechanics of handing a conversation from an agent to a person without losing context, see how chatbot human handoff should work. Shopify stores can install Ernest from the App Store and run the simulation against their own ticket history before deciding anything.