Summary of "Битрикс24 Day – большая презентация Нового Битрикс24. 14 мая 2026"

Business-focused summary (Bitrix24 Day 2026: “New Bitrix24”, 14 May 2026)

Core business thesis: AI shifts work from “tools” to “agents + automation”

The presenters frame AI as a step-change in how companies execute work: from searching/typing to getting decisions and execution done via agents, with communication, tasks, CRM, and even development increasingly automated.

1) Key strategic frameworks / operating playbooks emphasized

No formal OKRs/SWOT were stated, but several operating concepts and process patterns were repeatedly used:

Agent cycle (execution loop)
- Agent perceives the situation → thinks what to do → uses tools (search/code/actions) → checks the result → repeats until the goal is met.
Decomposition → orchestration
- Large tasks are broken into sub-tasks handled by multiple agents, then composed into a final deliverable (e.g., presentations/articles).
“One request → one outcome” task philosophy
- A task should converge to a single result; AI enriches tasks from chat/audio/video inputs.

2) Product strategy & operational changes in Bitrix24 (big releases)

A) Collaboration: new AI-first communication layer in Messenger

Chat becomes the primary collaboration interface (text + files + voice/video + reactions).
Chat folders
- Mobile now; later web/desktop.
- Used to organize conversation contexts.
Bitrix24 Syn (calls + recording)
- One-click joining.
- Cloud recording automatically shared.
Bitrix24 GPT follow-up after calls
- Automatically analyzes calls and generates:
  - soft-skill feedback
  - meeting objectives / actions
- Emphasis: meeting knowledge should not be “lost” after the call.

Concrete operational KPI claimed

~1,000,000 stickers per month sent internally on average (signal for emotional/chat engagement).

B) “Chats with external participants” (deal-making without granting full access)

External users can be invited into company chats via link/phone/email without Bitrix24 registration/authorization.
Access is scoped to conversation content: history/files/audio/video.

C) Tasks & projects: AI generates work artifacts from conversation

AI audio/video tasks
- User sends a voice/video instruction in chat.
- Bitrix GPT identifies the task and converts it into a structured task:
  - description
  - performers
  - checklist
Project AI (new “AI projects”)
- Centralizes fragmented project communication into one workspace:
  - project chat
  - files, tasks, meetings, calendar
- Includes multiple project agents (described as):
  - Agent 1: weekly info collection + publishing progress/analytics
  - Agent 2: personal assistant per employee (creates the optimal next set of tasks)
  - Agent 3 (implied): context-aware agent that uses full project context to answer specifically
Migration rule
- Existing groups/projects/collaboration are automatically converted into new AI projects, preserving data.

D) Knowledge Base (AI-ready collaborative documentation)

New Bitrix24 Knowledge Base with:
- modern block editor (text + tables + images + docs)
- role-based access per article/section
- collaborative editing inside Bitrix24
Positioned as an AI input asset for downstream automation.

E) Mobile operations: Check-in 2.0 + geolocation + mobile-first task plans

Designed for field/shift workers (presented as “up to 80% of employees” may not use a computer).

Check-in 2.0 capabilities

Start-of-day “emotion-enabled” check-in
Shift attendance metrics for managers (shifts worked / absences)
Geocoordinates verification
Multiple check-ins per worksite/object with result attachments
Work plans & daily reports generated by Bitrix GPT:
- plans appear in the employee’s chat
- daily summaries sent to the employee
- roll up to manager reports

F) Bitrix GPT 55: performance + multimodal agent behavior

Claim: Bitrix GPT runs on “Contours of Russia” for low latency.
Adoption metric: >88% of employees on paid tariffs use Bitrix GPT in daily work.
Subscription model:
- Bitrix GPT “unlimited” + 5,000+ apps from the marketplace.

Bitrix GPT 55 “agent” features

Thinking mode: shows reasoning process + more detailed answers
Online mode: if it lacks an answer, it goes online to collect and synthesize
Multimodal: images + documents/PDFs/tables
Long-context: understands tasks and related CRM/work artifacts

Rollout timeline

Bitrix GPT 55 agent mode available tomorrow on all portals
“Agency mode” rollout:
- by end of May: all portals receive agent mode
- by end of July: complete multimodal agent rollout

3) Wipecoding / Wipecode platform (the biggest operational shift)

Presented as enabling employees to create real working applications without traditional development bottlenecks.

How it works (operational model)

User generates an app by giving:
- a key
- documentation
- what they want to build
AI uses:
- API “methods” and Bitrix24 data access permissions
- a platform-specific deployment mechanism
Deployment is designed to be safe and fast (minutes).

Security & governance mechanisms emphasized

Blackhole servers
- isolated/invisible servers not directly reachable from the internet
- authorization-scoped access to needed data/apps
Admin controls:
- access scopes
- white/black lists
- restrict model access and time-based rules

Economic model emphasized

Pricing:
- pay per minute of deployment/usage
- “day-night rates” (server sleeps/wakes automatically; example mentioned: 29 days)

Platform adoption metrics (speaker)

1,583 users
~15,000 portals
440 applications created
~500–605 keys/servers (subtitle counts partially garbled; latest mentioned: 49 new servers added just yesterday)
Claim: active commercial usage by clients/partners; “incredible reviews”.

4) CRM & Digital Workplaces: “CRM AI” + speech analytics + agent-led sales operations

CRM AI: automate deal creation & enrichment from calls/chats/emails

Key automation stages

CRM setup using Bitrix GPT Suggests fields, stages, funnel, kanban transitions.
Call processing
- call → speech-to-text → extract deal fields → auto-fill CRM
- reduces manual time; example claim: up to 15 minutes of manager time saved per call
Chat processing
- extracts deal info and next actions
- captures “promises” into the deal so managers don’t re-listen/re-read calls
Spam protection
- AI identifies spam calls/chats/emails
- adds contacts to CRM exceptions to prevent deal creation

Business messaging

“Less effort, fewer people needed” to reach results via automation.

Speech Analytics 2.0 (automated script management)

Problem addressed

Managers previously maintained sales scripts manually (holidays, new offers, multiple business areas).

New approach

AI selects and updates the “right script” for each call type and manager performance.
Evaluates:
- compliance with script
- soft skills
- recommendations for improvement
Output delivered into messenger as bot digests / operational control.

Operational KPI claimed

AI scenarios save up to 40% of managers’ time
Presented as significant relief due to script maintenance complexity (usage cited as “third place”).

Repeat sales automation: conversion & NPS use case

Repeat sales conversion metrics

Current repeat-sales conversion: >26%, growing
Company audits:
- 48% in one company
- 56% in another
Target guidance:
- aim to bring repeated deals closed to ~50% (implied KPI goal)

NPS system built via Wipecoding

Custom forms collect customer NPS after purchase/service delivery.
If assessment is bad:
- automatically connect a manager
- triggers message + follow-up workflow
Engagement/conversion claims:
- conversion started around 4%
- later average 8.8% of voters
- “more than 9,000 responses” processed

Experimentation / optimization

AI optimizes dispatch time:
- earlier assumption: “evening”
- tested and found best sending time for that client segment: 10:00 am
Result: increased response frequency; “conversion rates growing year-on-year”.

5) Online Booking: service resource reservation + AI call scripts

Online booking is positioned for service businesses such as:

beauty salons, clinics, car services, rentals, and similar providers.

End-to-end operational playbook

Booking grid blocks:
- reserve specialists/resources per client
- prevents double booking
Notifications chain (real-time):
1. confirmation request the day before
2. morning reminder, with rationale that clients can change plans in the morning rather than 1 hour before
3. no-show handling: immediate message asking whether to wait/reschedule
4. if no response: manager contacts client and rebooks an alternative slot
AI “AI call script” for live rescheduling:
- agent calls client → checks availability → reschedules instantly

KPI/time-to-serve claim

If there’s no quick response, average client defection to a competitor happens within 10–15 minutes.

6) Concrete actionable recommendations (what companies should do)

Adopt agent-first workflows
- orchestrate work in chat/voice/video so tasks can be auto-generated.
Centralize project context
- stop scattering comms/files across channels; use AI projects to keep context for accurate agent answers.
Use the knowledge base as an AI training asset
- publish internal process docs with access rights so AI answers reliably.
Enable AI in CRM operations
- auto-fill deals from calls/chats; let AI handle spam filtering and promise tracking.
Use Wipecoding for internal automation
- create internal mini-apps (dashboards, calculators, workflow automations) instead of waiting for dev cycles.
Govern with security controls
- use blackhole servers + access scoping + admin policies for safe employee-generated apps.
Optimize sales operations with repeat sales + NPS
- measure repeat-deal conversion and feedback loops; personalize outreach timing and content with AI.

Presenters / sources mentioned

Maxim Zrodovsky (restaurateur/entrepreneur; spoke about learning/using Wipecoding)
Andrey Karpaty (credited with coining “wipe coding / wipecoding”)
Sasha Vartanyan (prepared “Wipecoding for humanities students” course)
Yury (referenced as “talk more about the platform”; likely co-presenter)
Andrey (mentioned in context of “Come on, Andrey, talk more…”)
Additional named entity/source: Claude and Codex (model leaders mentioned as current approaches)
“Wipecode / Wipecoding agents platform” and Bitrix24 leadership team (multiple speakers; names not consistently clear due to subtitle quality)

Share this summary

Is the summary off?

If you think the summary is inaccurate, you can reprocess it with the latest model.

Summarize another video