XBASE or a decision-tree chatbot · 2026
XBASE is for people who need to understand what the customer wrote, the way they wrote it, and keep the team in control: the AI agent helps visitors in the website chat, the AI suggests replies on WhatsApp and WhatsApp conversations stay in the pipeline.
A decision-tree chatbot is the bot that offers numbered options ("type 1 for…") and takes the customer down a fixed path. It is predictable and easy to check.
It is a customer service bot built like a menu: each customer answer picks a branch, and each branch has a text written in advance. It works well when there are few questions and they are always the same. When the customer writes something nobody anticipated, the bot goes back to the menu or hands off to a person, depending on how it was designed.
XBASE is a conversational CRM with AI. It brings official WhatsApp and Instagram Direct into one inbox, puts each conversation in a pipeline stage and keeps every customer in the CRM. Our expert team builds the automations and AI agents with you.
What changes
In the website chat, the agent understands whatever the customer types, with no "type 1".
Catalog, prices and policies, as PDF, Word, CSV or text.
The person answering sees the whole conversation and the customer profile. In the website chat, only the agent replies.
It writes the agent's instructions, goals and rules with you and tests before going live.
Side by side
A decision tree is predictable, uses no AI and always answers the same way. XBASE trades part of that predictability for the ability to understand open-ended messages, with rules, testing and the team close by.
| Capability | Decision-tree chatbot | XBASE |
|---|---|---|
| How it answers | ||
| Main product | Menu bot with fixed options | Conversational CRM with AI |
| DifferentiatorUnderstands open-ended messages | Recognizes the expected options | In the website chat. On WhatsApp, the AI suggests to the team, with the assistant set up with the expert team |
| Where the answer comes from | The text written for each option | The company's files |
| Instructions and goals | The tree design | With version history |
| Safety rules | Not needed: it only says what was written | Rules you turn on that filter profanity, email addresses, phone numbers and sensitive numbers |
| Same answer every time | Varies with the conversation; that is why there is testing before publishing | |
| Buttons and lists on WhatsApp | It is its format | Saved for reuse |
| Customer's language | Only the languages someone wrote | In the website chat |
| Team and records | ||
| The team takes over with the history | Depends on the tool | On WhatsApp, the team replies with the history and the conversation owner; in the website chat, only the agent replies |
| DifferentiatorPipeline and customer profile | Depends on the tool | |
| Score for each conversation | Depends on the tool | Per goal, with evaluation turned on |
| Improvements | Someone redesigns the tree | With evaluation turned on, suggestions drawn from the conversations, which you approve |
| Cost | ||
| Use of AI | None | The AI works on website replies and suggestions |
| Plan | Depends on the vendor | Monthly plan based on team size |
Main product
Understands open-ended messages
Where the answer comes from
Instructions and goals
Safety rules
Same answer every time
Buttons and lists on WhatsApp
Customer's language
The team takes over with the history
Pipeline and customer profile
Score for each conversation
Improvements
Use of AI
Plan
Legenddoes itdoes notpartial or depends on the tool
"Differentiator" marks the rows XBASE was built to win.
Checked on 26/09/2026 against the product in production. The chatbot column describes the menu-bot format, not a specific product.
The difference that matters
The tree waits for a number. The agent reads the sentence.
In the tree, each answer is written again. In the agent, it comes from the catalog and policies you already have.
The tree only says what someone wrote for each option. The agent follows the instructions you wrote and the safety rules you turn on.
What it costs
The tree's price depends on the vendor. XBASE plans are below.
Uses no AI. The price depends on the vendor.
Meta's per-message fees, on the card registered with Meta.
Prices and limits checked on 11/05/2026. Payment by credit card.
See plansTo decide
Choose XBASE if
Stay with the tree if
For each type of operation
The tree
Opening hours, address and a contact: the menu handles it.
XBASE
In the website chat, the agent answers with the catalog you keep up to date in its files.
XBASE
Each conversation in a stage, with the owner the team sets, and the AI suggesting replies.
To be fair
The tree always answers the same way and is easy to check. The agent can make mistakes: that is why you test before publishing, turn on safety rules and, with evaluation turned on, approve the suggested adjustments.
The tree uses no AI.
On WhatsApp, the XBASE AI suggests and the person sends. The agent answers on its own in the website chat.
How to switch
The questions that end in "invalid option" or "talk to an agent" show what the agent still needs to learn.
The option texts become agent files, and the expert team writes the instructions with you.
Talk to the agent before putting it to work on the website. Buttons and lists stay where they help.
Questions from people deciding
Some do. This page compares with the menu bot, with fixed options. If yours already uses AI, compare by what gets recorded: pipeline, profile, owner and proposal.
It can happen. That is why it uses the files you choose, follows the safety rules you turn on and is tested before going live. With evaluation turned on, it also gets suggested adjustments that only apply with your approval.
By monthly plan, paid by credit card, based on team size. Meta's per-message fees are separate. Prices and limits checked on 11/05/2026.
See plansThe expert team writes the instructions, goals and rules with you, organizes the files and tests before going live.
Yes. Messages with buttons and option lists are still available, saved for reuse.
On WhatsApp, the AI suggests the reply and the person sends it. In the website chat, the agent answers.
This comparison was written by XBASE and checked on 26/09/2026 against the product in production. The decision-tree chatbot column describes the menu-bot format with fixed options; since we don't cite a specific product, there is no third-party source.
From the first message to the proposal.
Create the account, teach the agent with your files and build the pipeline with our team.
Create account
