Planfix

Planfix connects as a channel: your AI Employee replies to customers with comments in tasks, and Planfix delivers the reply to the messenger the request came from. The connection consists of an API token, a channel in AI Textura, and an automated scenario in Planfix.

How it works

Planfix isn’t a messenger. A customer writes to you in Telegram, WhatsApp, by email, or in your website chat, and Planfix turns the request into a task and stores the conversation in its comments. When an employee replies with a comment, Planfix sends it back to the customer in the same channel.

Your AI Employee fits into this setup like a regular Planfix employee:

1. The customer writes

   The message arrives in the Planfix task as a new comment.

2. Planfix calls AI Textura

   An automated scenario sends the comment to your channel’s webhook URL.

3. The AI replies

   AI Textura adds a reply comment to the task.

4. Planfix delivers the reply

   The customer gets the reply where they wrote: in Telegram, WhatsApp, or by email.

You don’t need to connect messengers separately. All channels already set up in your Planfix automatically become available to your AI Employee through a single Planfix channel.

---

What you need

What	Where to get it
Planfix account with administrator rights	Your Planfix: you need the API Access and Automated Scenarios sections
Account domain	Your full Planfix address, including the zone: acc.planfix.ru or acc.planfix.com
API token	Created in Step 1
Bot employee in Planfix	A regular employee that tasks for the AI will be assigned to. You can use an existing one

Connecting takes three steps and about 15 minutes. The order matters: the webhook URL only appears after you save the channel, and you can’t set up the scenario without it.

---

Step 1. Create an API token in Planfix

1. Open the API section

   In Planfix, go to Account Management → API Access → REST API.

2. Create a new key

   Create a new API key and give it a clear name, for example “AI Textura.”

3. Grant permissions (scope)

   The token needs at least these permissions:

   Permission	Why
   comment_add	Post replies as comments
   contact_readonly	Identify which customer the request came from
   file_add	Attach files
   file_readonly	Read customer attachments
   task_update	Reassign the task on handoff to an operator
   user_readonly	Get the list of employees and groups for the Assign To field

4. Copy the token

   Save it somewhere temporarily; you’ll need it in the next step.

The API token grants full access to your Planfix account. Don’t send it via messengers or store it in shared documents. Once saved in AI Textura, the token is hidden and never shown again.

---

Step 2. Create a channel in AI Textura

1. Open the AI Employee

   Go to AI Employees → select an AI Employee → Channel tab → Connect New subtab.

2. Select Planfix

   Find Planfix in the platform list and click Create.

3. Fill in the two required fields

   • Domain: the full account address, including the zone: acc.planfix.ru. You can paste it with https://; the extra part is removed automatically.
   • API Token: the key from Step 1.
4. Click Connect

   The channel is saved, and AI Textura generates a unique, secret webhook URL for it.

5. Copy the webhook URL

   Open the channel you just created on the Connected tab. In the Webhook URL field, click the copy button. You’ll need this address in Step 3.

The Webhook URL field stays empty until the channel is saved, and that’s expected. The address is generated once when the channel is created and doesn’t change afterward.

The webhook URL works like a password. Anyone who knows it can send messages to your AI Employee on behalf of customers. Don’t publish it.

---

Step 3. Set up an automated scenario in Planfix

The scenario is what makes Planfix call AI Textura on every new customer message.

1. Open scenarios

   In Planfix, go to Account Management → Automated Scenarios and click New Scenario.

2. Set the event

   Event that triggers the scenario: Comment added and task matches conditions.

3. Set the conditions

   Add two conditions and combine them with 1 AND 2 logic:

   1. Comment author is a contact, so the bot only replies to customers, not to colleagues’ comments.
   2. Assignee = your bot employee, so the bot only works in tasks assigned to it.
4. Add the Send HTTP Request action

   In the Perform the following operations section, select the Send HTTP Request action and fill in:

   • Method: POST
   • URL: the webhook URL you copied in Step 2
   • Type: application/json
   • Request body: switch to the PARAMETERS tab (not “Text”)
   • Authorization: No authorization
5. Add the parameters

   Add six parameters. The names on the left must match exactly, character for character:

   Parameter	Value
   taskId	{{Задача.Номер}}
   commentId	{{Комментарий.Идентификатор}}
   text	{{Комментарий.Текст}}
   contactId	{{Задача.Постановщик.Номер}}
   contactName	{{Задача.Постановщик.ФИО}}
   fileIds	{{Комментарий.Прикреплённые файлы.Идентификаторы}}

6. Save the scenario

   In the Change notifications block, leave Notify selected and save.

For contactId, choose the requester’s Number, not “Identifier.” “Identifier” returns a different, technical number, and with it the reply won’t reach the customer’s messenger.

To send every new task straight to the AI, set the bot employee as the default assignee in the task template. Otherwise, you’ll have to set the assignee manually, and the scenario won’t fire until you do.

---

Additional fields: context from the task

Besides the six required parameters, you can pass any task fields to your AI Employee, such as city, contract number, plan, or request topic. It receives them along with the message and takes them into account in its reply.

Add more rows to the same parameter list. You make up the name yourself, and it becomes the label the AI sees:

Parameter	Value	What the AI sees
Request topic	{{Задача.Название}}	Request topic: Invoice not received
City	{{Задача.Поле.Город}}	City: Kazan
Contract number	{{Задача.Поле.Договор}}	Contract number: DG-7788

Name parameters in plain language, since the AI reads these exact labels. Contract number is clearer than contract_no.

Limits: no more than 15 additional fields, names up to 60 characters, values up to 300 characters. Empty fields aren’t sent.

---

Channel settings

Open the channel on the Connected tab to change its settings.

Setting	What it does
Domain	Your Planfix account address, including the zone
API Token	The access key. Hidden after saving; fill it in only if you want to replace it with a new one
Webhook URL	The address for the Planfix scenario. Read-only
Assign To	The employee or group the task is reassigned to on handoff to an operator. If empty, the bot simply removes itself from the task
Send Reply as Note	AI replies become hidden: Planfix employees can see them, the customer can’t
Allow Handoff to Operator	Lets the AI hand the request over to a human employee on its own. Off by default
When to Hand Off to Operator	Your plain-language description of when to bring in a human

Note mode

Turn on Send Reply as Note if you want to review what the AI replies before showing it to the customer. Replies appear in the task as hidden comments, marked with a lock and visible only to your employees.

It’s a convenient, risk-free way to test a new AI Employee on real requests. Once you’re happy with the replies, turn off the toggle, and the AI starts replying to customers directly.

---

Handoff to an operator

If you turn on Allow Handoff to Operator, the AI can bring in a human employee on its own. Here’s how it works:

1. The AI tells the customer it’s bringing in a colleague.
2. A hidden note with a short summary of the conversation is added to the task, so the employee doesn’t have to reread the whole thread.
3. The task is reassigned to whoever is set in the Assign To field.
4. The bot stops replying: the scenario no longer fires, because the bot is no longer the assignee.

When to hand off

Fill in the When to Hand Off to Operator field in plain words, no special syntax. For example:

If they ask about a refund or if the order total is over 100,000 ₽

If you leave the field empty, the default cases apply: the customer explicitly asks for a human, the AI can’t resolve the issue, or the conversation is about a refund or complaint.

Handoff to an operator is off by default, so turn it on deliberately. If the Assign To field is empty, the task is left without an assignee: the bot leaves it, but the system doesn’t assign a specific employee.

---

Attachments

If the scenario passes the fileIds parameter, your AI Employee receives the files the customer attached to the comment:

File type	What happens
Voice message	Transcribed to text; the AI replies based on its content
Image	The AI looks at the image and replies based on it
Other files	The AI receives the file name, so it can reply meaningfully rather than blindly

Up to 3 files per comment are processed, each up to 15 MB.

---

Check that everything works

1. Create a task

   In Planfix, create a task and set the bot employee as the assignee.

2. Write as a customer

   Add a comment from a contact, or ask a customer to write to the connected channel.

3. Wait for the reply

   Within a few seconds, a comment from the AI Employee appears in the task.

4. Check delivery

   Make sure the reply reached the customer in their messenger, unless note mode is on.

---

FAQ

Why is the API token field empty when I open the channel settings?

That’s by design. The token grants full access to your Planfix account, so once it’s saved, we never show it again, not even to you. The channel still works: the token is stored and used. Fill in the field only if you want to replace the key with a new one.

The Assign To list is empty and asks for the domain and token.

The employee list comes from Planfix, and the request needs the token, which, as described above, isn’t shown in the form. Paste the token into the field again, the list will load, then select an employee and save.

The AI doesn't reply to comments. What should I check?

In order: (1) the task assignee is the bot employee specified in the scenario conditions; (2) the comment was written by a contact, not an employee (the scenario ignores colleagues on purpose); (3) the scenario is enabled and the action has the correct webhook URL; (4) the channel in AI Textura is active and the AI Employee’s status is “Active”; (5) the AI Employee’s working schedule allows it to reply right now.

Planfix says "Scope denied" or "No access."

The token is missing permissions. Go back to Account Management → API Access and grant the key all six permissions: comment_add, contact_readonly, file_add, file_readonly, task_update, user_readonly.

The reply appears in the task but doesn't reach the customer.

The cause is almost always the contactId parameter: the scenario uses “Identifier” instead of “Number.” Change the value to {{Задача.Постановщик.Номер}}. The other possible cause is that Send Reply as Note is on, in which case only employees can see the reply.

Won't the bot start replying to itself?

No. First, the scenario condition “Comment author is a contact” filters out the bot’s comments. Second, Planfix doesn’t run scenarios for comments added via the API. Third, AI Textura discards duplicates by comment number. That’s three independent safeguards.

Can I connect several Planfix accounts?

Yes. Each account is a separate channel with its own domain, token, and webhook URL. Remember the rule: one channel is linked to one AI Employee.

What happens if I change the scenario or disable the channel?

A disabled channel stops receiving messages: Planfix keeps sending requests, but there will be no replies. If you need to hand conversations back to human employees for a while, turn off the AI Employee rather than the channel, so the conversation history is preserved.

What’s next

Channels

All the ways to connect customers to your AI Employee

Bitrix24

Another system that connects via webhook

amoCRM

CRM connection via OAuth

Notifications

An alert for you when a customer asks for a human