Landbot Review

Landbot is a strong choice when you want to control how website and WhatsApp inquiries become qualified leads, while letting AI answer questions along the way. This Landbot review recommends it for a small business with a defined intake process and someone responsible for human follow-up; its Free plan's 100 monthly chats are a place to learn the builder, not an ongoing AI support allowance.

Verified against Landbot's live product pages, pricing and help documentation on 5 November 2026. This evaluation compares documented capabilities and a worked workflow; it does not claim an account trial or measured conversion results. The decision rests on control over qualification, channel behavior, what reaches a person, and what triggers additional billing.

What Landbot is

Landbot is a visual conversation builder that lets a business connect scripted questions, AI answers, saved customer information and human support in a web or WhatsApp chat. You decide which questions must be answered, which replies can come from your business content, and when the conversation should move to another system or a person. For a marketing lead, that means replacing a generic inquiry form with an intake process. For a support lead, it means answering repeat questions and collecting context before escalation. It is useful when that process needs deliberate steps; it is less compelling when you only need a short contact form that already works.

Who should choose Landbot, and who should skip it

Landbot fits a small business whose inquiries need qualification before someone can act: a home-services company checking service area, an agency screening project requests, or a training business matching prospective students to courses. Its value is the quality of the information arriving with the inquiry. A chat that produces more messages but leaves the receptionist asking the same questions has not solved the problem.

The useful shortlist is about the job you need to improve:

These are buying priorities, not claims that only one product has a particular feature. All need a trial against the same customer request before you move a live channel.

Tidio: start here when the inbox is the problem

Tidio, a customer-service platform combining a Help Desk with its Lyro AI Agent, is the more natural starting point for a shop juggling order questions, live replies and unresolved tickets. Its current product includes ticketing, conversation assignment and ecommerce order handling. That makes a support-led evaluation sensible before you spend time designing a separate intake flow.

The tradeoff is operational fit: a capable inbox does not decide your qualification criteria or prove that a prospect reaches the right salesperson. If the problem is incomplete quote requests, compare the actual intake sequence rather than choosing on help-desk breadth. The Tidio alternatives guide expands the shortlist by support job.

Chatbase: start here when questions lead the conversation

Chatbase, an AI customer-experience platform built around connected knowledge and agent instructions, deserves a look when visitors mostly arrive with their own questions. Its current product includes Data Sources, Guardrails, Procedures and Actions, a testing Playground, human Helpdesk features and WhatsApp deployment. It is broader than a document-answering widget.

The constraint is that uploading information does not establish a reliable business process. You still need to verify what the agent captures, which actions it can take and how a person receives the result. Prefer Landbot when seeing and editing the qualification path is the main buying criterion. Use the Chatbase alternatives guide when knowledge handling and support integration are the larger decision.

The visual builder: make the qualification rules visible

Landbot's visual builder is most useful when an answer changes what should happen next. A local installation business should ask about the service area before asking a visitor to describe a project in detail. A repair inquiry should reach a different destination from a new installation quote. Those are business rules worth putting on a canvas where the person responsible for leads can inspect them.

Landbot describes message and question blocks, conditional paths, variables that store answers, formulas and reusable Bricks, which package a sequence for reuse. Its Copilot can draft an initial flow from a description, while templates provide another starting point. These features are documented on the chatbot platform page.

For the installation business, begin with the destination: an inquiry that a coordinator can accept, redirect or decline. Then work backward to the information needed for that decision. Service type and location belong early. Decorative questions and a long welcome message do not earn their place merely because the builder can display them.

Use fixed choices when the business has fixed categories. Leave room for a visitor to explain an unusual request. A branch that says “something else” is useful only if it leads somewhere a person can handle; otherwise it disguises a dead end as flexibility.

The maintenance advantage of a visible flow is accountability. The marketer can point to the rule that rejects an unsupported location or routes an existing customer to support. The maintenance burden is equally concrete: someone must update that rule when the service area or ownership changes. No-code removes a programming task; it does not remove the decision-making work.

Put the web conversation where the decision happens

Landbot supports a website widget, popup, embedded section or full-page conversation, plus appearance customization and flow analytics. Its website chatbot page describes those options.

For a quote-request page, an embedded conversation is the clearest candidate because the visitor is already trying to submit an inquiry. A widget makes more sense when questions arise while someone browses several service pages. A full-page experience can work for a dedicated campaign, but it should not force visitors through a conversation just to find basic business information.

Judge the change by usable inquiries and completed next steps. Watch where visitors leave, then read the question at that point. If a field is unnecessary before a coordinator can respond, remove it. If it is essential, explain why you need it. More engagement with the bot is not automatically more business.

Landbot AI: answer questions without surrendering the flow

Landbot AI works best as an informed part of the conversation, with clear boundaries around eligibility and commitments. The public AI Agent page describes answers grounded in business content and handoff with captured information and conversation history. That is useful for the questions a form cannot comfortably handle, such as whether a service includes a site visit or how to prepare for an appointment.

The important distinction is between explaining a policy and making a commitment. A home-services assistant can explain the published survey process. It should not invent a final quote or promise an appointment that has not been confirmed. Keep those decisions in the business rules, booking system or human review path.

The AI feature names describe different jobs

The AI Agent block is the part that talks to customers. Build it with AI creates its setup from a description; Edit it with AI revises existing Agent Instructions. The Knowledge Base holds business information, Store Data defines saved Fields, Interactive components supply elements such as buttons, and Outputs end the AI conversation and return control to the connected flow. These names and behaviors come from the AI Agent setup documentation.

For the installation inquiry, put service descriptions and preparation guidance in the Knowledge Base. Put behavior in Agent Instructions: answer from those sources, collect the required details, and hand off requests for a binding quote. Define matching Fields in Store Data. A field's existence alone does not instruct the agent to collect it.

Connect the “needs a person” Output to Human takeover. Connect a completed qualification Output to the next required business action. The AI Agent does not know what that next block does, so an exit condition should identify when to stop, not promise that a downstream email or booking has already succeeded.

This division makes the workflow reviewable: the flow owns required steps, the AI explains, and the connected system confirms actions. A friendly answer cannot stand in for a saved lead or a completed appointment.

Prepare knowledge that the agent can read

Landbot's documented Knowledge Base is text-only. A policy uploaded as an image, or a crucial instruction shown only inside a diagram, is not usable evidence for the agent. Open the imported source and inspect the text the agent received. Rewrite missing information as clear text before relying on it.

URL sources can refresh automatically on a selected schedule. Assign someone to own that schedule and the source page. If the business changes its service terms but leaves an older uploaded document in place, a polished answer can still be wrong.

A useful acceptance check is an ordinary service question, an ambiguous request, and a request the business must decline. Read the replies and inspect the saved information. You are checking whether the process holds together, not whether the assistant sounds impressive.

WhatsApp: turn a lead form into a qualification conversation

Landbot's WhatsApp channel is a strong reason to shortlist it when customers already use that channel to ask about your service. Landbot provides an official WhatsApp Business API connection, an opt-in widget, campaign tools, a test environment and a Team Inbox, described on its WhatsApp product page.

Consider a home-services business whose existing form collects a name, phone number and vague request. The coordinator then has to establish the location, type of work and desired next step by calling back. The proposed replacement is a short web entry that lets the visitor continue in WhatsApp, where the bot gathers the information needed for a useful human reply.

This is a worked design based on the documented building blocks, not a claim that this particular flow was deployed or produced a conversion lift.

1. 1

   Start with an explicit channel choice

   Offer a clear “Continue on WhatsApp” route from the inquiry page. Landbot documents a Click-to-WhatsApp option for a web bot in Growth tools. Tell the visitor what will happen after the click, and keep an ordinary contact route available for someone who does not want to use WhatsApp.

2. 2

   Make the WhatsApp entry self-contained

   Create the WhatsApp flow and define its opening question around the request: a new quote, an existing job or another question. Do not assume clicking away from a web conversation has automatically transferred every answer or matched the customer record. Confirm the details needed to continue, or deliberately connect a verified record through your integration.

3. 3

   Qualify with rules the coordinator can explain

   Ask for the service area and type of work, store the answers, and route unsupported requests to a useful explanation. Keep the qualification criteria explicit. Use WhatsApp-compatible reply choices, while allowing the visitor to describe an exception or ask for a person.

4. 4

   Use AI for the questions between the fields

   Let the AI Agent explain what a survey involves or what information the coordinator needs. Give it the current service guidance and instructions about what it must not promise. Then return to the qualification flow when the visitor is ready, or exit to a human when the request falls outside the approved information.

5. 5

   Save the request before handing it over

   Send the required details to the customer record and notify the responsible person. Put actions that must happen for every inquiry before Human takeover, because a successful takeover stops the bot. Give an unavailable team a separate follow-up path that confirms how the customer can be reached.

6. 6

   Test the conversation on a phone

   Use the WhatsApp testing option, then check a supported request, an unsupported location, a customer changing their answer, and a request for a person when nobody is available. Inspect the destination record as well as the transcript. A successful-looking chat with missing saved details is an unfinished workflow.

Landbot's WhatsApp building guide documents Click-to-WhatsApp, testing, reply options and Global Keywords. A keyword such as “Agent” can provide an escape to human support. It also warns that some web features are not fully compatible with WhatsApp even when they appear in the builder. Test the actual channel before copying a web design into it.

The continuation rule matters after the inquiry. Landbot describes a 24-hour window from the customer's last message; later proactive contact requires the appropriate opt-in and an approved Message Template. Collecting a phone number in an ordinary form is not a substitute for the WhatsApp consent flow described in those docs. Design follow-up while designing intake, so the coordinator is not left with a lead they cannot contact through the intended route.

For measurement, separate the question “Did this bot collect a usable request?” from “What happened across this WhatsApp number?” Landbot documents channel-level Metrics for the number and bot-level analysis for an individual bot. If several flows serve the same number, the individual bot view does not tell the whole customer story.

Integrations and human handoff: finish the job outside the bot

Landbot's integrations matter when the conversation creates work that someone must complete. A transcript is evidence of a conversation; a correctly populated customer record is something a coordinator can use. The integration catalog lists connections for HubSpot, a customer relationship management system, Google Sheets, Airtable, Calendly, Slack and Mailchimp, alongside automation options such as Zapier and webhooks.

For the home-services inquiry, decide which system owns the request. Map service type, location, contact details and the requested next step to its fields. Decide what should happen if the person already exists there. A native connection reduces setup effort, but your field names and ownership rules remain your responsibility.

Keep the success message tied to the action you can verify. If the conversation has collected a request but the receiving system has not confirmed it, “your request has been submitted” may be premature. Design a failure path that retains the information and tells a person what needs attention. This is workflow advice, not a promise that every connection provides the same retry or error-handling behavior.

For a straightforward supported connection, a marketer may be able to maintain the setup. For a custom endpoint, authentication, response handling and failures can require technical help. “No-code builder” does not guarantee that an unusual back-office system becomes a no-code integration.

Human takeover needs an online and an offline design

Landbot's Human takeover block assigns a conversation to an available teammate and stops the bot; if nobody is available, it follows the block's red output. The Human takeover documentation makes the consequence explicit: leaving that output unconnected can leave the visitor without an answer.

Add the block from Inbox & Builder tools, choose who can receive the chat, and connect the unavailable path to a useful message and contact collection. Put required notifications and record updates before takeover. Once a person receives the conversation, blocks after takeover do not run.

This changes the design for a small office whose coordinator also answers calls. “Busy” and “Offline” do not count as available for new takeover assignments. A reassuring “connecting you” message is therefore insufficient: the next path must explain what happens when the coordinator cannot respond.

Also review Settings > Preferences > Assignation for conversations already assigned when a teammate goes offline. Landbot documents options to retain, unassign or reassign those chats. Pick the behavior that matches how your office covers absences, and have the person responsible for follow-up inspect the unassigned inbox.

The practical payoff is fewer repeated questions and fewer abandoned requests. The handoff is complete when the person can understand the request, find its saved details and act. The existence of a live-chat feature alone does not establish that outcome.

Landbot pricing: the plan boundary that matters

Landbot's Professional WA configuration is the relevant starting point for the website-and-WhatsApp workflow above. Starter is a web pilot option, but it stops below own-number WhatsApp, premium CRM integrations and branding removal. The figures below are the USD amounts on Landbot's live pricing page, before tax; annual figures are monthly equivalents billed annually.

• Free: $0; 100 chats/month, 1 seat, no AI chats; WhatsApp testing with Landbot's number.
• Starter: $36/month annually or $45 monthly; 500 chats/month, 100 AI chats and 2 seats.
• Professional: $88/month annually or $110 monthly; 2,500 chats/month, 300 AI chats and 3 seats; WhatsApp costs extra.
• Professional WA: $176/month annually; 2,500 chats/month, 500 AI chats, 3 seats and 1 WhatsApp number.
• Business: custom pricing and chat/AI volume; from 5 seats and 1 WhatsApp number included.

Those are allowances, not an all-inclusive bill. Landbot meters AI chats separately, including conversations that use AI only for part of the exchange. It lists extra AI chats at $0.10, regular overage at $0.05 per chat billed in 500-chat packs, and extra seats at $25/month. WhatsApp message charges can add Landbot and Meta fees.

The decision is therefore not simply whether the included chat total looks large enough. Check how many inquiries will enter AI, who needs inbox access, and how follow-up messages are charged. Request the applicable message-fee schedule before budgeting a campaign; the page's USD view includes some fee cells in euros, so treating the headline subscription as a final invoice would be misleading.

Landbot's limitations that change the decision

Landbot's main limitations appear where an attractive conversation meets an operational requirement. These are the ones that should change your trial, plan choice or shortlist.

Web and WhatsApp are different experiences. The WhatsApp guide explicitly warns about feature compatibility. A web bot that looks right is not proof that the same choices, media or input behavior work correctly on a phone. If the business cannot maintain channel-specific checks, keep the first release narrower.

The Knowledge Base cannot read image-only information. A visual service-area map or a scanned policy can contain the answer your business cares about while supplying no usable text. Convert essential information into readable text and inspect the imported source. If your support knowledge is mostly pictures, source preparation becomes part of the project.

Build it with AI can replace work you already refined. The setup documentation says it overwrites the entire Agent Setup. Edit it with AI is the feature for revising existing Agent Instructions. Preserve your reviewed instructions and field definitions before regenerating, and inspect the result before publishing changes.

An AI exit does not prove a business action happened. The AI Agent has no knowledge of the next connected block. Likewise, a successful Human takeover prevents later bot blocks from running. Put compulsory record updates where every relevant path reaches them, and avoid wording that promises an unconfirmed action. This execution-order issue matters more than how naturally the bot phrases its replies.

The AI allowance is a separate constraint. A conversation can use ordinary qualification steps and still enter the AI meter. If most visitors ask an open question, plan around that behavior rather than assuming the regular chat allowance covers the whole experience.

A support feature still needs a support owner. The red output, teammate availability and reassignment settings determine what happens when the office is busy. Landbot can provide the paths; it cannot make an unattended inbox into a staffed service. Skip it if buying the tool is being used to postpone deciding who follows up.

Landbot review verdict: buy it for a process you can name

Landbot earns a place on the shortlist when you need a designed qualification path across web and WhatsApp, with AI answering between controlled steps. It is especially sensible for a small business whose staff repeatedly ask the same intake questions before they can quote, book or redirect a request.

Choose Tidio first when the immediate problem is managing a support queue and ecommerce conversations. Give Chatbase equal attention when most of the value comes from answering varied questions from connected knowledge. Keep the existing form when it collects enough information and a chatbot would merely add friction.

The decision flips on whether the workflow reaches a useful next action. For the first pilot, take an existing inquiry and require the bot to capture the necessary details, save them in the intended system, answer only supported questions, and handle both available and unavailable staff. Check the saved Fields in Analyze and read the AI Analysis transcript alongside the destination record. Read the transcript together with the receiving record. If the coordinator still has to reconstruct the request, fix the flow before widening access.

On Monday, have the marketing lead and the person answering inquiries agree on that acceptance check. Build the smallest journey that passes it, then use observed drop-offs and incomplete requests to decide what to change. That is a defensible reason to adopt Landbot; a more animated contact form is not.

Frequently asked questions

Is Landbot free?

Yes. Landbot has an ongoing Free plan, but it does not include AI chats. The separate paid-feature trial lasts 14 days and moves to Free afterward without requiring a card. Use the trial to evaluate AI; do not treat its access as the permanent Free allowance.

Can I start building my bot before my WhatsApp API access is approved?

Yes. Landbot's WhatsApp page describes a closed test environment for building and trying the bot with your phone before approval. That lets you check the conversation; it does not mean your business channel is already approved for customers.

Landbot alternatives

Shortlist Tidio when a support workspace is the priority and Chatbase when the experience is led by knowledge-based AI answers. Compare each against the same inquiry, saved-data requirement and human handoff, rather than assuming a shared channel name means equivalent behavior.

If you want the inquiry flow, integrations and human handoff designed around your existing systems, see AI customer service.