Skip to content
ChatterfyHelp Center
HomeGeneralCRMAI ManagerTrackerPlatformsFAQUpdate log
Main
Bot Settings
Text Recognition
Bot Settings
General — basic bot settings
Delays — bot delays
Notifications — Telegram Notifications for the Team
Basic setup after connection
Webhooks & API — receiving data from external sources
Custom Fields — custom fields
Sale Status — sales funnel stages
Automation & AI
CRM defaults
Flow Settings
AI Settings
Steps
Communication
Changes
Terms
Tracking
Additional
Messenger
Client card
Chat data
Events
Users
Mailings
Analytics
  1. CRM
  2. Bot Settings
  3. Text Recognition

Text Recognition

AI recognizes data in messages and saves it to Custom Fields. Related materials: What is Text Recognition Text Recognition is a Chatterfy AI feature that analyzes user text messages, identifies the ne

Updated June 21, 20262 min
AI recognizes data in messages and saves it to Custom Fields.
Related materials:
AI Recognition — deep dive
AI Recognition — processing media with AI AI Recognition is a module in Automation & AI: AI reads attachments (voice, photos, video note, screenshots) and turns them into text and data for the flow. W
Custom Fields — custom fields
Custom Fields (Bot Settings → Custom Fields) — a section for creating additional fields in the customer card. Standard fields already exist in Chatterfy: name, username, status, and Telegram ID. But o
Bot Variables in AI Prompts
How to reuse links, identifiers, and constants in “AI Manager — overview” — , step , and via the {{variables.name}} macro. Shortening long URLs and common patterns. Section reference — . Here, we cove

What is Text Recognition

Text Recognition is a Chatterfy AI feature that analyzes user text messages, identifies the necessary information, and automatically writes it into Custom Fields.
Simply put:
the user types something;
AI determines if there is any relevant data in the message;
if data is found, it saves it to the field.

Examples

Below are typical recognition use cases:
User message: `I deposited 10000 INR` → record the deposit amount
`my email is test@gmail.com` → save the email
`telegram: test.username` → save the username
`My ID is 928311` → save the ID
`Deposit was around 250 dollars` → save the amount

Important

This functionality only works on **Ai Reply** steps.

Where to find Text Recognition

The section is located in: **Bot Settings → Text Recognition**
Each rule consists of the following elements:
Action — AI action
Target field — Custom Field to save the result
Example — an example of the expected value
Prompt — instructions for the AI: what to look for and how to process it

How Text Recognition works

The user sends a message.
The AI analyzes the text.
The AI checks the Text Recognition rules.
If information is found: it extracts the value and writes it to the selected Custom Field.

Action

Currently, the main mode is Write. The AI writes the recognized value into the selected field.

Target field

The field where the recognition result will be saved. Commonly used:
deposit
email
username
country
manager
telegram
amount
domain
user_id

Example

Examples for the AI:
deposit → 10000
username → test.username
ID → 928311

Prompt

The main instruction for the AI: what to look for, how to interpret the message, what to save, and how to transform data.

Use specific instructions

Bad: `Recognize user data.`
Good: `If the user provides an email address, extract only the domain part after @ and save it into the domain field.`

Always add an Example

Providing an example significantly improves recognition stability.

Use separate fields

Do not store different data types in a single field: `user_info` is better replaced by `email`, `domain`, `deposit`, `country`.

AI saves the incorrect value

If the model retrieves an incorrect result: check the `Example`, Prompt accuracy, and the correctness of the Target field.
What to do:
clarify the data format
add more context
separate the rules

AI does not save anything

If no data is saved: verify that the field exists, check the selected Target field, and confirm the data format sent by the user.
What to do:
optimize the Prompt
test with a real conversation
check multiple examples

Preparing data for AI

Recognized values can be used in:
AI Reply
AI Routings
Send Message
Webhooks
CRM filters

Was this article helpful?

BackBot SettingsNextGeneral — базовые настройки бота

On this page

  • What is Text Recognition
  • Examples
  • Where to find Text Recognition
  • How Text Recognition works
  • Preparing data for AI