Integrating ASI Biont with Bitrix24 and 1C-Bitrix: No-Code CRM Automation — Real Case Study and Scenarios

Introduction

Every day, sales managers drown in routine: manually entering leads into the CRM, updating deal statuses, sending the same emails, reporting on completed tasks. According to HubSpot, sales teams spend only 33% of their time directly communicating with clients; everything else is administration and internal processes. When the CRM is Bitrix24 or 1C-Bitrix, the complexity of configuration adds to this routine: robots and business processes require a programmer, and the API documentation is not always clear to a beginner.

But what if automation could be described with words in a chat, without delving into technical details? This is exactly the problem solved by the AI agent ASI Biont. It connects to Bitrix24 via REST API and allows you to perform any operations with CRM data using natural language. Instead of writing PHP scripts or configuring webhooks, you give the agent an API key and simply ask: “Create a lead from the email”, “Select all deals in the status ‘Awaiting Payment’ and send a report to Telegram”. The agent itself forms HTTP requests to Bitrix and processes the responses.

What is ASI Biont integration with Bitrix24?

Bitrix24 is a CRM system popular in Russia and the CIS, which also includes tasks, documents, telephony, and other tools for managing a business. It has a powerful REST API documented on the official 1C-Bitrix developer portal. Through this API, you can create leads and contacts, modify deals, get reports, work with tasks and the sales funnel. ASI Biont uses this exact open API.

Integration in this case is not a ready-made module that needs to be downloaded from a marketplace. It is a dialogue with an AI agent that gets access to your Bitrix24 account. You don't write code and don't need to understand REST; you simply explain to the agent what needs to be done, and it turns that into API commands itself.

The traditional approach to integration looks like this:

1. You find a developer to study the Bitrix API and write scripts.
2. Then these scripts need to be hosted on a server or in the cloud.
3. When logic changes, you have to go back to the programmer.

With ASI Biont, everything is different. The user creates a webhook in Bitrix24, copies the link with the access key, and passes it to the agent in the chat. Then all the magic happens inside: the AI agent itself analyzes the API, checks permissions, and starts working. There is no need to wait for developers to add support for a new service — it is enough for the service to have an open API. The key principle of ASI Biont: connecting to any service via API without custom development.

Case Study: The Company “Logistic-Soft”

To make it clear, consider a fictional but typical example. The company “Logistic-Soft” sells software for transport companies. Managers work in Bitrix24, with a standard funnel: leads → deals → contracts. A sales department of three people performs many operations manually every day.

How it was before implementation

• Every request from the website comes to the general inbox; the manager has to open the email, manually create a lead in the CRM, fill in contact details, and then attach the file from the website.
• When a client rescheduled a meeting, the manager manually searched for the deal, changed fields, and set a new reminder in the calendar.
• The weekly report for the manager was compiled in Excel: data was exported from various statuses, grouped, and charts were made. This took about 2 hours every Friday.
• Duplicates often appeared in the database: the same client could be created twice if they first called by phone and then left a request on the website.

How ASI Biont was implemented

In the Bitrix24 settings, the administrator created a webhook with permissions for leads, contacts, deals, and tasks. Then, in the chat with ASI Biont, they wrote: “Use this API key to work with Bitrix24: ...”. The agent instantly checked access and displayed the list of available methods. Then the manager described the tasks, one by one.

Here are examples of commands that were launched:

• “Check the mail at sales@logistic-soft.ru every 15 minutes. If an email comes from a new sender, create a lead, fill in the fields “Name”, “Company”, “Email”, and set a task to call in 10 minutes. If the sender is already in the database, just create a deal from the contact.”
• “Every day at 9:00, send to the common chat the list of deals that have had no activity for more than 5 days and mention the manager.”
• “On Fridays at 18:00, upload funnel statistics: the number of deals at each stage, the amount, the number of new leads, and send the report to the owner in Telegram.”

The agent independently wrote code that accessed the Bitrix24 API. Managers communicated with it through the common chat and clarified details, for example: “Send not the entire list, but only overdue tasks, and exclude tasks with the tag “cold”.” After several iterations, the scenarios became ideal.

Results after a month

We compared the company's metrics before and after implementing the agent. We present the results in
the table below. For each metric, we recorded the average values over the last month before the implementation and the last month after.

The most striking change was in the speed of the first response. Previously, if a request came in at 8:00 PM, the manager would only see it the next morning. Now the agent creates a lead, sends a message to the chat, and sets a reminder for a call within 10 minutes – even at night. The manager opens the chat, sees the context, and calls the client back in the morning with all the details already in the CRM.

Another important result was the reduction of duplication. The agent checks the database for matching email or phone before creating a new lead. If a contact already exists, it simply adds a deal to that contact and notifies the manager. This has almost completely eliminated the chaos that happened when a client called and then sent a request via the website.

The weekly report also changed its nature. Instead of spending two hours exporting and formatting, the manager now gets a ready-made summary in Telegram every Friday at 18:00. The message contains a chart with the funnel stages, the total amount in work, and a list of deals that have been stuck for more than five days. The owner says he now checks the report from his phone while on the way home – and can immediately ask questions in the chat.

Of course, there were adjustments along the way. The agent initially created a task for every single incoming email, including newsletters. After two days, the managers asked to ignore messages with the “unsubscribe” link. Also, the agent sometimes confused the “company” field when the client wrote their name instead of the organization. We fixed this by adding a rule: if the sender’s domain is not a free email service, take the domain as the company name; otherwise, leave the field empty.

In the end, the implementation took five days: one day for the setup and API connection, two days for testing and refining the scenarios, and two days of parallel work with the old manual process. After that, the agent was left to work independently. The company’s managers now spend 15–20% less time on routine operations, and the owner estimates that the agent pays for itself in two months just by reducing missed leads.

If you have a similar routine in your CRM, you don't need to write complex scripts. Try describing your task to a business agent in ordinary language – you'll be surprised how many of these processes can be automated in one evening.Before you launch such an agent, make sure you have a clear idea of what should happen when something goes wrong. In our case, we set up a special chat where the agent writes every time it is unsure about a step, and a manager approves or corrects the action. During the first week, this happened a few times per day; by the end of the month, it was almost never. That safety net made the team comfortable enough to trust the automation.

It is also worth thinking about who owns the agent. At first, it lived under the administrator's login, which caused confusion when it created or modified records. We moved it to a dedicated service account with limited rights – it can read contacts and deals, create tasks, and send messages, but cannot delete anything or change the payment data. That calmed everyone down.

One more thing we learned: the agent is not a one-time setup. As the company grows, you will add new fields, new statuses, new rules. Budget time every month or two to review what the agent is doing, look at its logs, and adjust the instructions. It is literally a conversation – you can tell it in plain language that it should stop acting on old deals or change the reminder time, and it updates its behaviour accordingly. No developer needed.

If you decide to try this, start with a single process that annoys you the most. It can be lead qualification, follow-ups, or report generation. Run the old and the new process in parallel for a couple of days. Compare the results, listen to your managers, and then let the agent take over. In my experience, the hesitation disappears very quickly once people see that the boring work is simply gone.