Sergey Pedchenko, Head of Technology Development at Yandex 360, answers questions from the IXBT.pro analytical department.
What global goals were set for the team when planning Yandex 360 updates? How do they correlate with current market trends?
Our key goal in developing the strategy is to provide businesses with complete freedom of choice and predictability. We see that the market is moving in three main directions: increased security requirements, deeper integration of AI into everyday workflows, and the need for flexible deployment models. Our updates are a direct response to these requests. We focused not on disparate functions, but on creating a holistic virtual workspace with cross-service interaction, where ease of use, security, and powerful tools for collaboration and automation are combined with the ability to choose between cloud, hybrid, and on-premises models. In addition, it is important for us that clients receive a unified set of capabilities, so we set ourselves the task of ensuring that our on-premises solutions do not differ in functionality from cloud solutions. We want companies to choose a deployment model not based on limitations, but solely on internal necessity: in some areas, it is simply impossible to use cloud solutions.
Name 3–5 recent key changes that you consider breakthrough in Yandex 360. What makes them unique compared to previous releases?
Here, we should immediately mention neural networks in Yandex 360 services. Recently, we introduced Alice Pro, an AI assistant that will gradually be integrated into all services of our virtual office. Alice Pro is already integrated into Mail. It combines information from emails and uploaded files into your personal knowledge base.
The AI theme does not end there. Now, in Tables, AI-based functions are available to help you cope with tasks faster. And importantly, we were among the first in the world to do this. And we didn't just add AI functionality, but deeply integrated these capabilities into the service structure. Thanks to this, neural networks can analyze, structure data, translate texts in cells into another language, compress them into short theses, and more. And all without real chances of hallucinations, as the neural network uses RAG technology in its work. It allows you to work exclusively with user data, and not rely on the internal data of the neural network.
It is also worth noting that our on-premises solutions will not differ in functionality from SaaS. This is, in fact, an extremely important announcement for companies that, for one reason or another, cannot use cloud solutions in their work. Previously, such companies had to choose: either the maximum functionality of the cloud service, or limited, but with local installation. Now this question no longer arises: the choice is made based on internal or legislative requirements. But here, too, it is important to understand that more and more companies are choosing the third way - hybrid. It combines the two previous ones: where local data storage is required, you can use on-premises, and where the requirements are not so critical - SaaS.
Communication for any company is one of the key tools for growth. And here, too, we offer new solutions. Previously, we announced federations - the ability to combine several companies into one information space, thanks to which employees from different companies can work together in Yandex 360 services: schedule meetings, correspond, call each other, and quickly find each other in the contacts of Mail, Messenger, and Calendar.
We continue to develop the Telemost service, which this year has already received neural network features, and in the near future will expand its capabilities by introducing messenger functionality. This will make it a universal communication tool for both business and personal tasks.
In general, we should not forget about small but important changes. Thus, shared drives have appeared in the virtual office - these are special storage locations where you can store files, work together with colleagues, and search for the necessary information. Shared drives can be used by employees who have received access from the organization's administrator. The administrator can create, delete shared drives, clear the recycle bin, and configure access.
How does the new Yandex 360 strategy combine AI, on-premises solutions, and security into a single ecosystem? What is the fundamental difference from competitors' approaches?
Our strategy is based on the principle of "one platform — different models." The virtual office has unified interfaces, functions, logic. The difference is in the location of data storage and processing.
Security is a cross-cutting element of the entire system: unified policies, audit, two-factor authentication, Secure client, which we will launch next year, thereby increasing the level of data security. We offer not a set of disparate tools, but a holistic, flexible, and secure environment for work. But we do not forget about functionality either. Thus, AI can work both in our cloud (maximum simplicity and speed of implementation) and in the on-premises environment, although the latter requires more resources.
We hear the market's requests for ease of use of the virtual office and increasing its efficiency, and not creating functions for the sake of functions, and we satisfy them.
How does the strategy affect the TCO (total cost of ownership) for business? Compare the costs of cloud, hybrid, and on-premises models, taking into account AI functions.
Our strategy gives businesses a tool for managing TCO by choosing the optimal model. If the key factors are speed and minimizing initial investments, then the cloud SaaS model will be the most profitable. You get all updates, including new AI functions, in particular, Alice Pro, automatically, without the cost of infrastructure and its maintenance. It is in the cloud that the launch of AI is most optimal in terms of resources.
When requirements for control and data security come to the fore, then on-premises is indispensable. Here, TCO is different: capital investments in equipment and deployment are required, but you get full control over the life cycle and data. For AI on-premises, significant computing power is needed. The hybrid model, which we also support, allows you to flexibly distribute the load and costs, for example, leaving critical data on your side, and tools for collaboration and communication taking from the cloud. Thus, we do not impose one model, but allow you to choose an economically justified path.
What three directions of strategy development do you consider priorities for 2025–2026? How will they respond to the growing demands of business in the field of data sovereignty and AI automation?
If we highlight three key vectors, then it is, firstly, the deep development of the built-in AI assistant. We will expand the contextual understanding of Alice Pro, introducing it into even more work scenarios. She should become not just a function, but a real personal assistant who anticipates the employee's tasks.
Secondly, it is the deepening of security capabilities. We are actively working on strengthening built-in audit and control tools, as well as on integrations into ready-made information security platforms, such as Yandex Security Deck.
And thirdly, it is the movement towards a unified super-application for work. We see a demand for maximum convenience and seamlessness. Our goal is that the transition between mail, documents, chat, and video meetings is not just fast, but contextually linked, creating a truly holistic digital work environment.
How has the collaboration platform (Documents, Tables, Presentations) been updated? Have functions for synchronous teamwork in real time appeared?
Yes, this year we focused on working in large teams. Thus, more than 400 people can already use the document editor simultaneously, while each of their actions will be recorded and taken into account. Collaborative editing is useful in large-scale projects with a large number of participants: for example, when working on annual reports or corporate regulations.
To enable hundreds of users to work on a document together, Yandex 360 has created its own implementation of CRDT (Conflict-free Replicated Data Types) technology. It is used where data needs to be constantly synchronized, such as in online games, messengers, and note-taking services.
As for Presentations, we are preparing a major update for this service next year. Firstly, it will also switch to its own technological solution, and secondly, it will receive a number of important features, which we will talk about closer to the announcement.
What new project management tools have been added to Yandex 360? How do they integrate with existing services?
Our main project management tool is Yandex Tracker. We are constantly developing it and making it more convenient and efficient. For example, in September we completely changed the approach to the mobile version. The large-scale update focused on simplifying the interface and improving navigation. The ability to view task boards has been added. In addition, more attention was paid to integration with products that are part of the Yandex 360 virtual office. A separate menu has appeared on the main page for switching to other services. To access applications or open Mail, Documents, Calendar, or Telemost in a mobile browser, simply click on the special button at the bottom of the screen.
Changes have also affected comments. They can be sorted — for example, from old to new or vice versa, and long messages can now be scrolled. A light version of the editor is built in for editing text in discussions. It allows you to highlight key phrases, change the design, and also leave short reactions if there is no need to write a detailed answer.
To help users quickly find out about new messages, a notification section has appeared in the application. An icon in the form of a bell has been added to the bottom panel. There you can view notifications individually or mark several as read at once. This option reduces the likelihood of missing important events.
What AI technologies are integrated into Yandex 360 services? Give examples of automating routine tasks through AI.
We are implementing AI not as a separate "feature", but as an assistant that becomes part of familiar tools. Our goal is to take on as much routine as possible so that people can focus on meanings and solutions.
Specifically, AI in Mail analyzes user behavior patterns. Anonymized data includes, for example, the format of interaction with letters: which ones people open, which ones they mark as important, and which ones they delete. Based on more than 1000 parameters, the neural network creates a personalized digest — highlights the most important letters: from clients, colleagues, with approaching deadlines. This radically changes the morning ritual of parsing the mailbox — instead of an hour of studying letters, you see the essence in 5 minutes.
In the Documents editor, AI is built directly into the editor interface. It helps not only to correct errors, but also to rewrite the text in a different style, shorten it, make a plan, or even generate a draft based on theses. It's like a smart co-author at your fingertips.
One of the most popular scenarios is in Telemost. The neural network transcribes the conversation in real time, and after the meeting it gives a structured retelling with the allocation of decisions, deadlines and speakers. You don't need to rely on memory or take notes — the whole essence is recorded automatically.
For working with data, we have seriously pumped up Tables. Now you can work with them in natural language: ask to "show sales dynamics by region for the quarter" or "highlight the top 10 clients". The system itself will understand the request, build the necessary formula or diagram. This removes the barrier in the form of complex Excel functions.
And, of course, Alice Pro. This is already the next step — an assistant who can analyze the context. In Mail, she can, for example, independently find the details, the term of the contract or the final amount of the invoice in the chain of correspondence and instantly give you an answer. You can "feed" her a document — and she will highlight the essence, find hidden contradictions or overdue obligations. Now she is actively working with Mail, and the plans include her deep integration into all services of the ecosystem.
Thus, AI for us is not an abstract technology, but a concrete assistant for parsing mail, stenography of meetings, editing texts and analyzing data, which already saves our users hours every week.
What specific business tasks (besides routine operations) can the AI assistant solve in Yandex 360? Give examples of scenarios where it replaces or complements the work of employees.
It is able to enhance analytical and management functions. For example, conduct an analysis of large volumes of feedback from mail with the allocation of common topics. Or prepare preliminary reports based on data from Tables and discussions in the chat. That is, it becomes a "partner" who processes information, leaving the person to work with context, decision-making and communication.
Tell us about the scalability of solutions. How does Yandex 360 adapt to the needs of large corporations with thousands of employees?
The architecture was originally designed for scalability. For corporations, centralized administration through group policies is important, rather than management "by users". In addition, we provide deep integration with existing systems through standard APIs for process automation.
What new administration capabilities have IT specialists received? Has it become easier to manage access rights and security settings?
It has become much easier. There are advanced action logs, two-factor authentication and more flexible management of it, fine-tuning of access rights to Disks and meeting parameters. In development is a unified support center that will allow the IT department to more effectively process employee requests regarding Yandex 360.
How have the updates affected the speed of services? Have algorithms been optimized to reduce the load on the infrastructure?
For us, it is important that updates work for users and do not cause them inconvenience. Therefore, we not only add new features, but also constantly improve our technologies. Optimization is a continuous process. In addition, the transition to our own technological solutions allows us to achieve a huge advantage: we managed not only to fully integrate AI into a significant part of the services, but also, without losses, for example, to bring conferences in Telemost to a new level. Thus, it became possible to hold meetings for up to 1000 people. And in Documents, as I said earlier, now several hundred users can edit a document. Also, the service is able to process large files without any slowdowns.
What is the key difference between the "Protected Client" and standard encryption mechanisms? How does it close the risks that traditional DLP systems do not cover?
"Protected Client" is not a replacement for security tools, but an additional add-on to ensure a comprehensive approach to information security. It will be a single application for accessing all virtual office services, designed to provide maximum data protection when working with both corporate and personal devices of employees. It will be created based on technologies used by Yandex Browser for organizations.
The main task of the product is to ensure maximum data security. The client includes a built-in secure storage where all information is encrypted. Data loss prevention and environment control policies restrict the transfer of data to third-party applications. Thus, documents cannot be, for example, copied to a third-party messenger or intercepted through the device's file system.
The protected client supports the BYOD (Bring Your Own Device) model, allowing employees to securely work from personal devices without risk to corporate information. Administrators will be able to restrict access to services from other clients, directing users exclusively to a secure environment.
How are "Protected Client" and AI services integrated into existing IT landscapes (ERP, CRM, Active Directory)? Provide examples of seamless integration.
Standard APIs are used for integration. Thus, the Protected Client can be integrated into any customer's information security solution. We initially wanted not to overcomplicate the structure, but to carefully integrate into it to provide a new level of security.
What are the plans for the future? What services or functions does the team plan to develop in the upcoming releases of Yandex 360?
A large team is working on the Yandex 360 virtual office, looking at the product as a whole, and not concentrating on specific services. Therefore, the development of the product is synchronous, and not in leaps and bounds. We presented the main results and announcements at our own Yandex Connect conference this year. Firstly, we are going to integrate Alice Pro into our other services. Secondly, by the end of 2026, we will prepare on-premises versions of all our services. Thirdly, we will continue to develop our core services so that they grow and develop together with the business that uses them.
