BitrixGPT Instead of CoPilot: How Bitrix24's AI Assistant Changed in 2026

BitrixGPT instead of CoPilot: what happened to Bitrix24's AI assistant

In spring 2026, Bitrix24 began a gradual rebrand: the familiar CoPilot assistant is being renamed BitrixGPT. The change isn't happening everywhere at once — the new name is already visible in the messenger, tasks, and some tools, while other sections will get it later. But the new label is more than marketing: alongside the rename, the assistant picked up a noticeably expanded set of features, and that's what actually matters to businesses more than the name itself.

For companies already running Bitrix24 as their CRM and corporate portal, the update means the AI tools inside the system are no longer an "extra option for enthusiasts" — they're becoming a working layer that touches sales, documents, websites, and employees' everyday tasks.

What actually changed

Beyond the name, the assistant gained several notable additions.

CoPilot Pro roles

The assistant can now take on more than 30 roles — sales manager, copywriter, marketer, lawyer, and other specialists. Instead of one generic chatbot, a company gets a set of narrowly focused helpers, each tuned to a specific type of task and speaking the language of the relevant profession. That's useful for companies without a dedicated lawyer or marketer on staff but with a manager who occasionally needs that kind of expertise.

Generating websites through dialogue

One of the most practical updates is the ability to create and edit websites and landing pages through an ordinary conversation with an AI agent, without manually working the block builder. The user describes the desired result, and the system assembles a draft page that can then be fine-tuned by hand. For small businesses, this lowers the barrier to entry: a landing page for a promotion or a new product can now be put together by an employee with no web design skills.

Boost add-ons

Paid extensions called Boosts have appeared — they raise the limits and speed of specific features without requiring an upgrade to a pricier plan overall. For example, there are separate Boosts for speeding up document generation and for increasing the number of AI requests, in some cases multiplying the limit many times over. That changes the budgeting logic: instead of overpaying for an entire higher-tier feature set, a company can top up precisely the one thing that's actually hitting its limit.

The assistant is also now fully available in the mobile app at the same level as on desktop — previously, some CoPilot features only worked on a computer.

How Bitrix24's AI layer got here

CoPilot didn't launch as a standalone product — it arrived as a layer on top of existing modules: CRM, tasks, drive, websites. At first it offered narrow prompts: a short call summary, a draft reply to a client, auto-filling a deal card from correspondence. That format proved the assistant was in demand, but it also exposed a limitation — users had to decide for themselves which task, in which module, warranted calling the AI, and there was no unified scenario.

The move to BitrixGPT and a role-based system changes exactly that logic. Instead of scattered prompts, a company gets a set of ready-made "staff members" it can address with a specific request — treated as a specialist rather than a feature buried in the interface. That's closer to how major CRM platforms worldwide are structuring generative AI: not as a separate button, but as a layer woven into every workflow.

Why this matters for business in Uzbekistan

For local companies working with Bitrix24 through partners, the assistant's update is worth attention for three reasons.

• Less load on small teams. CoPilot Pro roles partially cover tasks that previously required a dedicated specialist or contractor — a draft commercial proposal, an initial contract review, copy variants for a mailing.
• Faster launch of marketing activities. Generating landing pages through dialogue cuts the time from idea to working page from days to hours, which stands out especially before seasonal promotions and sales.
• Flexible spending on AI features. The Boost model lets a company pay for exactly the volume of AI requests it actually needs instead of forcing the business onto a fixed plan.

That said, it's worth understanding that the assistant's expanded limits and roles are almost always tied to paid add-ons or a higher plan. A free request package exists in the cloud version, but it's meant for a trial run of the tool, not for the sustained workload of a sales or marketing department.

There's a fourth, less obvious reason that matters for companies serving clients in several languages. CoPilot Pro roles can be used to draft correspondence and materials directly in the needed language — Russian, Uzbek, or English — taking some of the load off staff who have to switch between languages throughout the day. It doesn't replace a human translator for official documents, but it noticeably speeds up routine correspondence and internal material prep.

What to watch out for when adopting the new features

Before expanding AI assistant use across a company, it's worth going through a few checks.

Access rights and customer data

The assistant works with data already sitting in the CRM — deal cards, correspondence, documents. Before actively using the roles, it's worth checking which employees have access to which parts of the portal, so the AI doesn't surface information that should stay off-limits to a particular user.

Verify results instead of trusting blindly

A contract draft, legal note, or commercial proposal generated by the "lawyer" or "sales" role is a draft, not a finished document. That's especially true for wording tied to local legislation and tax rules: AI models are trained on huge volumes of general data, but they don't guarantee accuracy on narrow, local-specific questions.

Check what's available in the on-premise edition

Many of the new AI features land first in the cloud version of Bitrix24, reaching the on-premise (boxed) edition later, tied to specific release updates. If a company runs the on-premise version, it's worth checking with the vendor which of these features are already available in the installed edition and which will arrive with the next planned update.

Set limits and budget deliberately

Before turning on a Boost permanently, it helps to review real usage statistics from a trial period — how many requests the team actually makes and which roles get used most. That way, a company only pays for what delivers measurable value instead of an unused buffer.

Train staff to phrase requests well

The quality of what a role produces depends heavily on how clearly the request is phrased. A short internal cheat sheet with examples of good and bad prompts is a simple step that saves the team hours of trial and error when rolling out new AI roles.

Bottom line

Renaming CoPilot to BitrixGPT is really a prompt to notice how much more functional Bitrix24's AI layer has become over the past few months: more than 30 assistant roles, website generation through dialogue, a flexible Boost model, and full functionality in the mobile app. For businesses in Uzbekistan, this is primarily a tool for speeding up routine tasks — from document drafts to promotional landing pages — not a replacement for specialists. To make the new capabilities useful rather than risky, it's important to check employee access rights in advance, treat AI output as a draft that needs review, and turn on paid add-ons deliberately, based on actual usage statistics rather than marketing promises.