Blog// AI agents

How the Agent Economy and AI-Disrupt PDLC Are Changing Software Development

At TechCommunity Fest 2026, Sber outlined a shift to the agent economy and AI-Disrupt PDLC, a development standard built on GigaCode, GitVerse, and GigaStudio.

09-10-2026 · 8 min read
How the Agent Economy and AI-Disrupt PDLC Are Changing Software Development

TechCommunity Fest 2026, Sber’s annual festival for engineering communities, took place in Moscow. Habr reports that this year’s program focused on how artificial intelligence is transforming software development. The discussions featured Kirill Menshov, Senior Vice President and Head of Technology at Sberbank; Sergey Ryabov, Senior Managing Director and Director of Sberbank’s AI Solutions Development Department; and Dmitry Sokolov, Head of Sberbank’s Corporate Architecture Department. The festival’s centerpiece was a foresight dialogue titled “Sber 2029: How We See the Bank’s Technological Future.”

The executives argued that sustaining business growth—Sber has joined the ranks of the world’s 150 largest companies—requires major changes to new products and services and adapting them to the agent economy. According to Habr, the agent economy will inevitably develop in Russia along with the rest of the world, and Sber itself must undergo an internal transformation toward agentic development. One part of that transformation is AI-Disrupt PDLC. For Sber, this approach goes beyond internal technology modernization: the company is turning AI-Disrupt PDLC from an internal efficiency tool into an industry standard open to the external market. It is built on Sber’s own technology products: GigaCode, GitVerse, and GigaStudio. Agentic development has already improved productivity inside Sber; the company is now extending that experience through partnerships.

Alongside the main topic, the article mentioned several unrelated notes: a report of an incident at a Yandex data center in Sasovo, a brief remark about the cost of code, advice to avoid consuming minerals unnecessarily, and a recommendation to work with the regulator to resolve an erroneous block by the TSPU. These items are not part of the discussion about the agent economy, but they appeared in the original material.

What is the agent economy, and why does it matter?

The agent economy is a model in which autonomous software agents carry out tasks, interact with other agents and people, and make decisions based on data and goals set by humans. In software development, this means that agents can take on some of the work involved in design, coding, testing, and deployment. Festival speakers said that moving toward agentic development is necessary to remain competitive: companies that implement these approaches successfully can release products faster and reduce the cost of routine work.

Sber emphasizes that the agent economy is already taking shape globally, and its development in Russia is only a matter of time. The company therefore views it not as an experiment, but as a necessary part of its future technology strategy. Adapting successfully to this model requires more than new tools; it also means changing the culture of engineering teams. The focus shifts from writing code to managing agents, defining their goals, and checking their results.

How Sber is implementing AI-Disrupt PDLC

AI-Disrupt PDLC is Sber’s internal methodology for integrating artificial intelligence throughout the product lifecycle, from idea through support. According to Habr, the festival announcement said Sber plans to take the methodology beyond its own organization and offer it as an industry standard. AI-Disrupt PDLC is based on three of the company’s own products:

  • GigaCode — a platform for AI-assisted code generation and review;
  • GitVerse — a collaboration system for code repositories, where agents can suggest changes and conduct code reviews;
  • GigaStudio — an environment for designing interfaces and user flows with automated assistants.

Sber says pilot use of AI-Disrupt PDLC inside the bank has already produced a measurable increase in productivity. The company is now actively building partnership programs to share this experience with external customers and suppliers. Organizations ready to adopt the standard would gain access to tools and methods previously available only inside Sber.

What this means for products and services

The shift to an agent economy changes how products are created. Instead of a linear process in which each stage depends on specialists doing everything manually, teams can build cycles in which agents automatically generate code drafts, suggest architectural improvements, run tests, and even prepare release notes. For end users, that could mean faster updates, fewer errors, and services that adapt more readily to changing market needs.

However, as participants in the foresight dialogue noted, these changes require a rethink of familiar roles. Engineers will need to spend more time configuring agents, defining their objectives, and checking the quality of their outputs. Without clear governance, the risk of incorrect or insecure code increases. Successful adoption of AI-Disrupt PDLC therefore depends on building monitoring, auditing, and feedback systems that keep people in control of the final result.

What this means for small businesses in Russia and the CIS: practical steps

For owners of shops, service companies, or small manufacturers, the key question is how to use these opportunities without maintaining a large development team. According to Habr, Sber is opening AI-Disrupt PDLC as a standard, creating the possibility of access to tools that were previously closed to external users. The following steps can be started now and are based only on facts in the source:

  1. Assess your current IT processes. Identify which development or support tasks for your website, mobile app, or internal systems take the most time and involve repetitive work—for example, writing standard functions, basic testing, or preparing release notes. These are the best candidates for delegating to agents.
  2. Explore Sber’s available products. Since GigaCode, GitVerse, and GigaStudio are named as the foundation of AI-Disrupt PDLC, ask Sber representatives about access programs for external organizations, such as pilot projects, trial periods, or partnership terms. The article does not say whether offers for small businesses are already available, so confirm this directly.
  3. Set up a pilot with clear goals. Choose one small module, such as a website contact form or a simple service-price calculator. Try generating its code with an AI tool, then have your team or an external auditor review the result. The goal is to measure time saved and identify possible code-quality problems.
  4. Introduce basic review rules. Even when an agent generates the code, establish a review process. Assign someone to check that it meets security and performance requirements and matches the business logic. This helps prevent risks from automated errors.
  5. Train the team in working with agents. Training can cover how to formulate tasks for AI, interpret its outputs, and refine prompts when needed. These skills reduce dependence on external consultants and speed up adoption.
  6. Follow updates to the standard. Since Sber presents AI-Disrupt PDLC as an industry standard, follow the company’s official channels—blogs, webinars, and technical documentation—to keep up with new tool versions and usage guidance.

These steps do not require major infrastructure investment, but they let you begin experimenting with agentic development today using the resources becoming available as Sber opens up its internal experience.

Conclusions: where the catch is and what to watch

The source emphasizes that the agent economy and related approaches have already improved productivity inside a large corporation, but extending them to the external market requires reliable management and control systems. The main risk is expecting complete automation without recognizing that agents still need clearly defined goals and human oversight. For small businesses, that means benefits will come only if they invest time in setting up feedback processes and training staff.

The article also does not specify what pricing or access terms GigaCode, GitVerse, and GigaStudio will offer to external users. Before starting a pilot, confirm the commercial details and any limits on usage volume or the types of tasks that can be delegated.

Overall, if agentic development is treated as a tool that complements rather than replaces human expertise, small companies can release updates faster, reduce routine errors, and focus on creating value for customers. Success will depend on building transparent review processes and on the team’s willingness to learn how to work with new kinds of assistants.

Frequently Asked Questions

What is agent economy in the context of development?

Agent economy implies a decentralized system where autonomous AI agents perform specialized tasks, interacting with each other to achieve the overall project goal. This enables automation of many development and management stages.

How does the AI-Disrupt PDLC standard affect the product lifecycle?

AI-Disrupt PDLC (Product Development Life Cycle) based on GigaCode, GitVerse and GigaStudio automates and accelerates all stages of the product lifecycle, from idea to deployment. This reduces time to market and improves the quality of the final product.

What key advantages does implementing GigaCode, GitVerse and GigaStudio provide?

GigaCode provides code generation and optimization, GitVerse manages versions and collaboration among AI agents, and GigaStudio offers an integrated environment for development and testing. Together they form a powerful ecosystem for automated development.

Based on: Habr — TechCommunity Fest explores how the agent economy will change products, services, and development

Analysis & review by
Alexander K.

Web development, digital advertising and marketing expert since 2012. Founder of AIFORSALE.

How we can help
Online storesMarketingAI agentsContact us
← All articles