An AI chatbot for ecommerce answers shoppers in website chat, WhatsApp or Instagram DMs, and the best ones do two jobs: they help customers choose and buy, and they check, change, return or refund orders afterwards. For a high-volume consumer brand, choose by whether one AI does both jobs on your customers' channels, what it may do to an order, and who builds and maintains it. Flowcall, Gorgias, Fin (formerly Intercom), Rep AI and Tidio describe both jobs in one AI; LimeChat sells them as separate suites, and Ada, eDesk and ManyChat lead with one.
Facts checked 30 September 2026. Every vendor fact below comes from that vendor's own documentation, product page or app listing, checked on 30 September 2026 unless another date is given, and linked. We make Flowcall. It's listed first because this list is written for high-volume consumer brands, the buyers it's built for.
Short answer
- Order, return and refund requests and product questions on WhatsApp, Instagram, Messenger, email, live chat and phone, with one AI set up and improved by the vendor's team: Flowcall.
- A Shopify-first brand that wants a helpdesk and an AI agent from one vendor: Gorgias.
- A Shopify brand already on Intercom: Fin for Ecommerce.
- A large, multilingual brand with an existing CRM or contact center: Ada.
- An India D2C brand whose customers shop and ask for help on WhatsApp: LimeChat.
- Mostly on-site selling, with support on the same assistant: Rep AI.
- Selling on Amazon, eBay or Walmart as well as your own store: eDesk.
- A small store answering mostly website chats: Tidio.
- Sales that start in Instagram comments and DMs: ManyChat.
What an ecommerce chatbot has to do: sell and support
Lists of ecommerce chatbots mix two kinds of product: shopping assistants that sell, and support bots that answer order questions. Some lists split them into separate tiers with different goals and measures. The split is useful for measuring, but it isn't how customers write.
The same shopper asks "does this come in medium?" on Monday and "where's my order?" on Thursday, and on WhatsApp both questions land in the same thread. If two bots answer those questions, the shopper meets two histories and your team runs two tools.
The two jobs need different things from the chatbot:
| Before purchase (selling) | After purchase (support) | |
|---|---|---|
| What shoppers ask | "Which serum for dry skin?", "Is the blue one in stock in M?", "Will it arrive by Friday?" | "Where's my order?", "Change my address", "It arrived broken", "Where's my refund?" |
| What the AI needs | Your live catalogue: products, variants, stock and prices | Order, courier and payment data, permission to change orders, and a rule for each change |
| What "done" means | The shopper finds the right product and buys it | The order problem is solved in your systems, not just answered |
| How it goes wrong | Recommends something sold out, or invents a feature | Refunds outside your policy, or shows one customer another customer's order |
| What to measure | Orders and order value from conversations | Requests resolved without the customer coming back |
A chatbot that answers only from your help pages can explain the returns policy but can't say where a customer's refund is. One that reads your catalogue but not your orders can sell but not support. The ones that do both read live data for each job and, after the sale, act on the order within rules you set: the idea behind conversational commerce. For what AI should do across the rest of your support operation, see our buyer's guide to AI customer service for ecommerce.
One shopper, one WhatsApp thread, two ways
On Monday evening, a customer messages a skincare brand on WhatsApp: "I need a sunscreen for oily skin that doesn't leave a white cast." On Thursday she asks where her order is. On Saturday it arrives with a cracked cap, and she sends a photo.
Two bots. The brand's shopping assistant lives on the website, so the WhatsApp bot answers Monday's message with a link to the sunscreen collection and she chooses on her own. Thursday's question gets a tracking link. Saturday's photo gets the returns policy and a form. The next Monday a human agent opens the store admin, finds the order, asks for the photo again because the form didn't attach it, and creates a replacement.
One AI. The same AI answers all three. On Monday it asks whether she prefers a gel or a cream, shows three matching sunscreens with prices and sends a link to buy. On Thursday it gives the courier's expected delivery date for her order. On Saturday it checks the photo against the damaged-item policy; the claim is inside the window and under the approval limit, so it creates the replacement and confirms it in the chat. Had a rule failed, a human agent would have received the conversation, the photo and the order.
The customer and the channel are the same in both versions. What changed is whether one AI could see the catalogue and the order, and act under the brand's rules.
This is how Flowcall's AI Agent handles it. Its AI shopping assistant finds, filters and compares products from your catalogue and sends them with images, prices and buy links. In India, it also takes payment in the chat with a payment link or UPI, India's instant bank payment system. The same AI, on the same WhatsApp number, checks the order, reads the damage photo and creates the replacement once your checks pass. We built selling and support into one AI because customers don't write to a sales department and then a support department; they write to one brand's number.
How we chose
We started from the products recommended by the pages that rank for "AI chatbot for ecommerce" and "ecommerce chatbot" in the US and India, and kept the ones a consumer brand's support team could run for both jobs, or for one job very well. We checked each against five questions, using only the vendor's own documentation, product pages and pricing page:
- Before purchase: can the AI search the live catalogue, recommend products and take the shopper to checkout?
- After purchase: can it look up this customer's order and change it (cancel, edit, return, refund), and under which rules?
- Where the AI itself works: a helpdesk can receive WhatsApp messages while its AI answers only website chat, so we list the AI's channels.
- Handover: does a human agent receive the conversation and the order when the AI stops?
- Who builds and maintains it: your team, or the vendor's?
We left out developer frameworks and B2B lead-generation bots; builders such as Botpress and Chatbase are in our list of AI chatbots for customer service. We didn't test the products hands-on, and we don't compare prices: each entry says how the vendor charges and links its pricing page. After Flowcall, the entries run from platforms for larger support teams to tools for small stores and social selling; it isn't a quality ranking.
AI chatbots for ecommerce compared
| Best for | Before purchase (vendor docs) | After purchase (vendor docs) | Where the AI works | Who builds it | |
|---|---|---|---|---|---|
| Flowcall | Consumer brands with high volumes of order, return and refund requests and product questions | Finds, filters and compares products; photo search; buy links in the chat, and payment links in India | Refunds, cancellations, exchanges, address changes and replacements once your rules pass | WhatsApp, Instagram DMs, Messenger, live chat, email; human agents take calls in the same workspace | Flowcall's team, with your approval |
| Gorgias | Shopify-first brands | Shopping assistant with selling-style and discount settings | Cancel, return, address change (actions you switch on) | Email, chat, SMS, Instagram DMs, Messenger, WhatsApp | Your team |
| Fin | Shopify brands on Intercom | Product questions, recommendations, upsells, guidance to checkout | Order tracking, returns, exchanges | Chat, email, WhatsApp, Instagram, Messenger, SMS, Slack, phone | Your team |
| Ada | Large multilingual brands with an existing CRM | Product questions, stock checks, size guides, prompts based on browsing | Order status, return labels, exchanges, refunds, store credit | Chat, email, SMS, voice, WhatsApp, in-app | Not stated |
| LimeChat | India D2C brands on WhatsApp | Growth AI agent, WhatsApp checkout, abandoned checkout, cash-on-delivery (COD) checks | Order tracking, returns, refunds, cancellations, complaints | WhatsApp; email, SMS and Instagram as add-ons; voice | Not stated |
| Rep AI | On-site selling on Shopify, WooCommerce or Salesforce Commerce Cloud | Prompts on hesitation and exit, product narrowing, upsells | Tracking, returns and exchanges, cancellations, address changes | Website chat, email, Instagram, Messenger, WhatsApp | Your team |
| eDesk | Marketplace sellers | Answers pre-sale questions | Order, return and cancellation questions with live order data | Marketplaces, email, Instagram, Facebook, WhatsApp, TikTok Shop | Your team |
| Tidio (Lyro) | Small stores, mostly website chat | Starts conversations on product pages, compares, adds to cart | Status, refunds, cancellations, order edits, address changes, returns | Chat, email, Instagram, WhatsApp | Your team |
| ManyChat | Social selling | Comment-to-DM replies, AI conversations, broadcasts | Not described in its help centre | Instagram, TikTok, Messenger, WhatsApp, Telegram, SMS, email | Your team |
The 9 best AI chatbots for ecommerce
1. Flowcall
Best for: Consumer brands with high volumes of order, delivery, return and refund requests, as well as product questions, across WhatsApp, Instagram, Messenger, email, live chat and phone, who want one AI to support and sell on the same number and want Flowcall's team to set it up and keep improving it with them. Pricing: Flowcall shares pricing on a call.
Before purchase. Customers describe what they want, and the AI shopping assistant finds, filters and compares products from your catalogue, with images, prices and buy links. It searches from a photo or screenshot, filters by size, colour and price and, in India, takes payment in the chat with payment links and UPI. Catalogues come in from Shopify, WooCommerce, Magento or files.
After purchase. Flowcall's AI Agent runs on the SOP Engine: it follows your written procedures for returns, refunds, exchanges, cancellations, cash on delivery and damaged products step by step, and your rules make each decision. Order, delivery and payment details come from your systems: Shopify natively, and couriers such as ClickPost, Shiprocket and Delhivery through configured connectors. Once the checks pass, it completes refunds, cancellations, exchanges, address changes and replacements. It reads the photos and invoices customers send, collects addresses in WhatsApp forms, and replies in the customer's language, with more than 100 languages supported for messaging.
Beyond the chat. On the same platform, WhatsApp Marketing sends order updates when an order ships, is delayed or is delivered, sends abandoned cart text messages, and confirms cash-on-delivery orders with a discount to pay online in the same message. A reply can start an AI Agent workflow.
When a human agent is needed. The AI creates a ticket in Flowcall's AI Helpdesk with the conversation, a summary, the customer's orders and what's left to do. If the same customer writes on another channel while the issue is open, it stays on that ticket, which is what omnichannel customer service means in practice. Human agents take calls from Exotel, Ameyo, Ozonetel or Contaque in the same workspace.
Who runs it. Flowcall's team sets up the workflows, catalogue and integrations with you, tests each workflow against simulated conversations before launch, and keeps improving it from quality findings; your team approves each change. Flowcall can also keep Freshdesk, Salesforce or Sprinklr as your system of record.
2. Gorgias
Best for: Shopify-first brands that want an ecommerce helpdesk and an AI agent that sells and supports, from one vendor, configured by their own team. Pricing: published; helpdesk plans by ticket volume, with AI Agent charged for the interactions it automates (Gorgias's pricing page; our Gorgias pricing guide explains the plans).
Gorgias's AI Agent works on email, chat, SMS, Instagram DMs, Facebook Messenger and WhatsApp (Gorgias: AI Agent explained, updated in September 2026). As a shopping assistant it helps shoppers find products and buy, with settings for selling style and discount strategy; Gorgias says it's trained on your Shopify store, greets visitors proactively and suggests bundles (Gorgias: Shopping Assistant). After a purchase, it takes the actions you switch on, such as cancelling an order, processing a return or updating a shipping address, and hands over when it isn't confident, the topic is sensitive or the shopper is frustrated.
Check. It's built around Shopify, with one AI Agent per connected Shopify store, so ask how it handles orders that also run through a separate order-management or warehouse system. Your team configures the actions and handover rules. See also Gorgias alternatives.
3. Fin for Ecommerce
Best for: Shopify brands already on Intercom that want one AI for product questions and order support. Pricing: published; charged per outcome, such as a resolution (Fin pricing; see our Intercom pricing guide).
Fin for Ecommerce answers product questions as customers browse, recommends products and narrows the catalogue, suggests upsells and guides shoppers to checkout; after purchase it handles order tracking, returns and exchanges in the same conversation (Fin for Ecommerce). Fin can be deployed on the Intercom Messenger, email, phone (Fin Voice), WhatsApp, Facebook Messenger, Instagram, SMS and Slack (Intercom: channels for Fin, checked 28 September 2026) and answers in more than 45 languages. Your team sets up its procedures and data connectors and tests them in simulations.
Check. Fin says Fin for Ecommerce is only for Shopify-powered businesses (not WooCommerce, Magento, BigCommerce or custom storefronts) and requires the Intercom helpdesk; it isn't sold on its own. Read how Fin counts a resolution: it includes conversations where the customer leaves without asking for more help (Intercom: Fin outcomes, checked 28 September 2026). See also Intercom alternatives.
4. Ada
Best for: Large brands, especially multilingual ones, that want an AI agent working over an existing CRM or contact-center platform. Pricing: not published; ask Ada for a quote.
Ada's AI agents work across chat, email, SMS, voice, WhatsApp and in-app messaging in 60 languages, and connect to Salesforce, Microsoft Dynamics, Zendesk, ServiceNow, Genesys, Twilio Flex and systems such as order management and product catalogues (Ada for ecommerce). After purchase, Ada describes order tracking, return labels, exchanges, refund confirmations and store credit. Before purchase, it covers product questions, stock checks and size guides, and can start a conversation based on what a shopper is doing on the website.
Check. Ada's ecommerce page leads with support, so test its product discovery on your catalogue. Its public pages don't say who builds and maintains the workflows. Ada counts an automated resolution when the AI fully addressed the issue without escalating to a human agent, so map that to your own measure.
5. LimeChat
Best for: India D2C brands whose customers shop, pay and ask for help on WhatsApp, and who want WhatsApp commerce and support from one vendor. Pricing: not published; set by suite (marketing or support), conversation volume and AI automation tier (LimeChat pricing).
LimeChat sells two suites. The WhatsApp Marketing and AI Growth suite includes one-click abandoned checkout, reorder and feedback messages, COD verification, a growth AI agent for sales assistance and checkout inside WhatsApp. The AI Support suite puts an AI agent on chat and voice for order tracking, returns, refunds, cancellations and complaints, with image recognition and handoff to live agents (LimeChat pricing). WhatsApp is the main channel, with email, SMS, Instagram and an omnichannel CRM as add-ons. It lists Shopify, WooCommerce, Magento and BigCommerce among its store integrations, with Indian couriers and payment providers such as Shiprocket, Delhivery, Razorpay and GoKwik (LimeChat integrations).
Check. Selling and support are separate suites, so ask which AI a customer meets on each question and whether the support agent sees what the growth agent sold. Channels beyond WhatsApp are add-ons. For a side-by-side, see Flowcall vs LimeChat.
6. Rep AI
Best for: Brands whose main goal is on-site selling on Shopify, WooCommerce or Salesforce Commerce Cloud, with support on the same assistant. Pricing: plan tiers, with a free trial (Rep AI pricing).
Rep AI engages shoppers proactively when they hesitate, compare products or look likely to leave; it narrows products from their answers, searches in natural language and times complementary recommendations. After purchase, it lists order tracking, starting returns and exchanges, cancellations and address changes, and routes to human support when needed (Rep AI). It works in website chat, email, Instagram and Facebook DMs and WhatsApp, has its own inbox, and integrates with helpdesks including Gorgias, Zendesk, Freshdesk, Kustomer and Salesforce Service Cloud (Rep AI integrations).
Check. Its pages describe selling in depth and each support action in a line, so run your returns and cancellation rules through it on real orders before you rely on it for support.
7. eDesk
Best for: Brands selling on Amazon, eBay, Walmart or other marketplaces as well as their own store. Pricing: published; per user, with the AI Agent charged per outcome it resolves (eDesk pricing).
eDesk's AI Agent identifies the type of query (return, cancellation, order status, troubleshooting, policy question or pre-sale question), retrieves the order, channel, customer history, tracking and product information, replies in your brand's tone and escalates to a human agent when it's uncertain (eDesk AI Agent). It connects more than 300 marketplaces and channels, including Amazon, eBay, Walmart, Mirakl, TikTok Shop, Instagram, Facebook, WhatsApp, Shopify, Magento, BigCommerce and WooCommerce.
Check. Its AI page describes answering more than acting, so ask which actions it completes on each marketplace and in your store. It answers pre-sale questions but doesn't describe product recommendations.
8. Tidio (Lyro)
Best for: Small and mid-sized stores whose support is mostly website chat, with some email, Instagram and WhatsApp. Pricing: published; Lyro is billed per conversation, with AI usage capped by plan (Tidio: Lyro; Tidio pricing).
Tidio puts chat, email, Instagram and WhatsApp in one inbox. Its Shopify App Store listing says Lyro acts within your rules on order status, refunds, cancellations and order edits, in Shopify or through an API. On the storefront, it starts conversations on product pages, compares products and adds them to the cart (Tidio listing, rated 4.8 from 1,345 reviews). Lyro Actions add address changes and return labels, built in a visual builder and tried in a test environment (Tidio: Lyro Actions). For other small-store options, see our round-up of Shopify customer service apps.
Check. No phone channel is listed. Lyro pauses when you reach your plan's AI limit, so check what customers see when it does, especially during a sale. If you're outgrowing it, see Tidio alternatives.
9. ManyChat
Best for: Brands whose sales start in Instagram comments and DMs, TikTok or WhatsApp broadcasts. Pricing: published; plans by active contacts per month, under a pricing model introduced in March 2026 (ManyChat: Pro plan).
ManyChat automates conversations on Instagram, TikTok, Facebook Messenger, WhatsApp, Telegram, SMS and email. Its Instagram comments trigger replies under a post or reel and follows up with a private DM (ManyChat: comments trigger), and its AI Step holds goal-based conversations from the business context you give it (ManyChat: AI Step). Paid plans cover selling through DMs and moving leads from Instagram or TikTok into WhatsApp.
Check. It's built for marketing conversations; its help centre documents flows, triggers and AI Steps rather than order lookups, returns or refunds, so plan how order questions reach your support team. On Instagram, a private reply to a comment doesn't open the 24-hour messaging window until the shopper replies or taps a button.
Also worth a look
- Bloomreach (Loomi conversational agent): an enterprise AI shopping agent that recommends products within your own merchandising rules. It doesn't handle support itself: when a shopper needs help, it hands the conversation and its full transcript to Sierra's AI agent (Bloomreach: Loomi Shopping Agent page, checked 2 October 2026).
- Zendesk AI agents and Freshdesk Omni: if your team already runs either helpdesk, start with its own AI agent. Both are support-first; we compare them in Flowcall vs Zendesk and Flowcall vs Freshdesk.
- Shopify-only chat apps: our list of the best chatbot for Shopify options covers lighter, self-serve apps.
- WhatsApp-only platforms: see our list of the best WhatsApp chatbot platforms.
How to choose: test both jobs in every demo
Feature lists look alike, and a demo goes well on questions the vendor picked. Choose from your own conversations instead.
- Sort last month's conversations into the two jobs. Before purchase: product questions, sizing, stock, delivery times, offers. After purchase: WISMO ("where is my order?"), cancellations, address changes, returns, refunds, damaged items. The balance tells you whether you need a shopping assistant, a support agent, or one AI for both.
- Check where those conversations happen. If most arrive on WhatsApp or Instagram, ask for the demo there, not in website chat. On WhatsApp, a business can reply freely only within 24 hours of the customer's last message; after that it needs an approved template (Meta: sending messages, checked 28 September 2026). An AI waiting for a photo or an address has to follow up inside that window. Flowcall's AI Agent collects addresses, photos and documents in WhatsApp forms inside the conversation, and when a customer goes quiet it reminds them what it's still waiting for and picks the workflow up where it left off when they reply.
- Run the same checks on every shortlisted product, on your busiest channel:
| Check | Job | What to do | A pass looks like |
|---|---|---|---|
| Search by description | Selling | Ask for a product the way customers do ("a sunscreen for oily skin, no white cast") | Real products from your catalogue with current prices; no invented products or features |
| Out of stock | Selling | Ask for a size or colour that has sold out | Says so, and offers the nearest alternatives as alternatives |
| Buying in the chat | Selling | Ask to buy the product it recommended | A buy or payment link for the right variant |
| Identity | Support | Ask "where's my order?" from a number or email that isn't on the order | Asks you to confirm who you are; never shows another customer's order |
| Live order data | Support | Ask about a real order in transit | This order's courier status and expected date, not just a tracking link |
| A rule | Support | Ask to cancel an order that has already shipped | Checks the status, doesn't cancel, and offers what your policy allows |
| Damage | Support | Send a photo of a broken item | Checks it against your policy, then creates the replacement or refund, or hands over |
| A limit | Support | Ask for a refund above what you'd let the AI approve | Stops and routes it for approval |
| Switching jobs | Both | After a product question, ask about an earlier order in the same chat | The same conversation carries on, with nothing asked twice |
| Handover | Both | Ask for a human agent | The human agent gets the conversation, the order and what's been done |
| Language | Both | Switch to your customers' second language halfway | It follows the switch without losing the order |
- Ask who builds and maintains it. New ranges, sale prices, a shorter return window: someone has to update the catalogue, the rules and the tests. Find out whether that's your team, the vendor's team or both, and how a change is tested before customers see it.
Measure the two jobs separately
Give each job its own measures before you sign:
- Selling: orders and order value from conversations, compared with similar sessions without a chat (shoppers who ask questions are already more likely to buy, so compare like with like), and a weekly list of products the AI couldn't find or explain.
- Support: requests resolved without a human agent and without the customer coming back about the same order within seven days; refunds or cancellations a reviewer would have refused; and handovers where the human agent had to ask for something the customer had already given.
Vendors count resolutions differently. Fin counts conversations where the customer leaves without asking for more help, eDesk charges per outcome when the AI answers or acts, and Ada counts conversations the AI addressed without escalating. Agree your own definition before a pilot, and read a sample of conversations every week.
Flowcall reports the order value that came from conversations alongside CSAT for the AI Agent and human agents, first response time, resolution and the AI automation rate, in pre-built and custom dashboards, so you can watch both jobs in one place. Reading a sample only catches what you happen to read. Flowcall's AI Auto QA audits AI and human conversations for wrong or made-up answers, loops, missed handoffs, and promises or actions the AI wasn't allowed to make, and quotes the conversation in each finding.
Which should you choose?
| If your priority is | Look at |
|---|---|
| Order, return and refund requests and product questions on WhatsApp, Instagram, Messenger, email, live chat and phone; one AI for both jobs; workflows built and improved with the vendor's team | Flowcall |
| A Shopify-first helpdesk and AI agent from one vendor, configured in-house | Gorgias |
| Staying on Intercom with a Shopify store | Fin for Ecommerce |
| Many languages over an existing CRM or contact center | Ada |
| WhatsApp marketing and support suites for an India D2C brand | LimeChat, or Flowcall |
| On-site selling first | Rep AI |
| Messages from Amazon, eBay or Walmart in the same place as your store's | eDesk |
| SMS as a main support channel | Gorgias, Fin or Ada |
| A small store answering mostly website chats | Tidio |
| Sales that start from Instagram comments | ManyChat |
For the whole support stack rather than the chatbot, see our guide to ecommerce customer service and our comparison of ecommerce help desk software.
If Flowcall sounds like the fit, book a demo and we'll show Flowcall's AI Agent recommending products and then handling an order, return or refund in the same WhatsApp conversation, the rule behind each decision, and what your team sees on handover. Qualifying brands get a free proof of concept on the service problem they most need solved, whether that's product recommendations, a returns workflow or quality checks across every conversation, built with their team and tested before anything goes live.
Frequently Asked Questions
Which AI chatbot is best for ecommerce?
It depends on your channels, your store platform and who will build the AI. For a high volume of order, delivery, return and refund requests, as well as product questions, on WhatsApp, Instagram, Messenger, email, live chat and phone, look at Flowcall. Shopify-first brands that want one vendor should look at Gorgias, and Shopify brands on Intercom at Fin for Ecommerce. Ada suits large multilingual brands, LimeChat WhatsApp-first India D2C brands, eDesk marketplace sellers and Tidio small stores.
What is an ecommerce chatbot?
An ecommerce chatbot is software that answers online shoppers automatically, in website chat or messaging apps. The better ones read your live catalogue to help customers choose and buy, and your order, courier and payment data to track, change, return or refund orders within your rules, handing the rest to a human agent. For the difference between answering and acting, see chatbot vs conversational AI, and for real conversations, our chatbot examples.
Can an AI chatbot recommend products?
Yes, when it's connected to your live catalogue. It should ask a question or two, show real products with current prices and stock, admit when nothing matches, and link to checkout. Test it on your own products, including sold-out sizes, before you trust its recommendations. A dedicated AI agents for ecommerce guide covers what these agents can do beyond recommending.
Can an ecommerce chatbot handle returns and refunds?
Yes, if it's connected to your store, order and payment systems and allowed to act. The rules matter most: the return window, the product's eligibility and previous claims should be checked against the order data before a refund runs, and amounts above a limit should go to an approver. Our guide to returns management covers the workflow.
Does an ecommerce chatbot work on WhatsApp?
Many do, but check that the AI itself works there, not just the inbox. WhatsApp allows free-form replies for 24 hours after the customer's last message and approved templates after that, and Meta charges for the templates you send and, since 1 October 2026, for replies after 1,000 free service messages a month per business phone number (see WhatsApp Business API pricing). For WhatsApp-specific options, see our guide to the WhatsApp chatbot for ecommerce.
How much does an AI chatbot for ecommerce cost?
Vendors charge per conversation (Tidio's Lyro), per outcome or resolution (Fin, eDesk's AI Agent), per automated interaction on top of ticket-based plans (Gorgias), by active contacts (ManyChat), by suite, volume and automation tier (LimeChat), or by quote (Ada). Some offer a trial or a few free conversations to start. Model the bill on your busiest sale week, not an average month. Flowcall shares pricing on a call.
