EN
+7 (917) 274-98-10Discuss a pilot

On-premise · inside your perimeter · any open model

Turnkey corporate AI. You keep the keys.

An AI assistant for every employee, running on your company’s servers. It answers from internal documents with links to sources and works in your CRM and 1C within the permissions you set. Your data, models, code and access stay yours.

First call: 30 minutes, online. We’ll sign an NDA before discussing any details.

since 2006
building software
100+
projects
50+
mobile apps
Company perimeter · works offline
Flatstack · Chatinside your perimeter

Summarize supply contract No. 214

Equipment supply, payment in two installments, 24-month warranty. Amount: RUB 4,800,000. Term: 30 business days. Penalty: 0.1% per day.

  • Clause 2.1
  • Clause 7.3

Trusted by

  • Premium Parking
  • Twiage
  • ScanCafe
  • Westwood Global Energy Group
  • Kid & Coe
  • Legally
  • Clubworx
  • VehoHealth
  • Substantial
  • RMCsoft
  • SmartHead
  • EdNavigator

What it looks like

One product for employees, managers and IT

These are screens of the platform, not mockups. We will show it live at the meeting — including on your sample documents.

An employee asks — the assistant answers from documents

  • An answer with a link to the clause of a policy or contract
  • No answer in the documents — it says so honestly
  • Sees only what the employee has access to
Web chat: an answer about business trips with links to clauses of the policy; clause 3.2 of the source document is open on the right
An answer with links to document clauses. A click on a link opens the clause itself.

The agent does the routine, a person approves what matters

  • The agent finds data in the CRM and prepares an action
  • A significant action waits for approval on the web or phone
  • The agent takes system access from the vault for the task only
Web chat: the CRM agent has prepared a task and waits for approval — a card with parameters and Approve and Reject buttons
The agent has prepared an action in the CRM. It will not run until an employee approves it.

IT and security manage everything in one back office

  • Users and roles from AD/LDAP, knowledge sources, agents
  • Access vault, audit log, policies
  • Models and GPU load, executive dashboards
Back office overview: active employees, questions, share of answers with sources, actions awaiting approval, secrets due for rotation, usage chart and a “needs attention” list
Overview for IT and security: usage, share of answers with sources, approvals, secrets due for rotation, GPU load.

Mobile app

iOS and Android: the assistant and approvals in your pocket

The same conversation history, answers with sources and approval of agent actions — from the phone. Distributed via corporate MDM; answers are not stored on the device after logout.

  • Mobile app: a summary of a supply contract with links to clauses 2.1, 7.3 and 9.2
    An answer with links to contract clauses
  • Mobile app: a push notification “CRM agent requests approval” over the conversation list
    A push when the agent needs a decision
  • Mobile app: the CRM agent action card “Change deal stage” with Approve and Reject buttons
    Approval in one tap
  • Mobile app: search across documents available to the employee and recent documents
    Documents available to the employee

Interface shown in Russian. The company, employees and documents on the screens are fictional.

Why your own AI

Three problems corporate AI solves

Access to knowledge

Employees spend hours searching policies and email threads and keep asking colleagues the same questions. The assistant answers in seconds and shows where each answer comes from.

Data control

Contracts, client data and internal materials must not be sent to public chatbots. The platform runs inside your perimeter, and answers respect document permissions.

AI governance

The business needs consistent rules, predictable costs and IT oversight. Instead of scattered subscriptions, you get one platform with an activity log and central administration.

Platform

One platform, three ways to work

We choose the model to match your tasks, language requirements and available hardware.

Web and mobile chat

Apps for the browser, iOS and Android. Shared chat history, file handling and access to agents.

API

Connects AI to portals, CRM and internal apps. Every answer comes with its sources.

Agents

Carry out tasks in company systems within the permissions you set: review calls, create CRM tasks, prepare documents, compile reports.

POST /v1/chat → 200 OK · answer + sources

Modules

What’s in the platform

Knowledge

The knowledge base connects to Confluence, file storage and CRM and re-indexes changes on its own. It answers with source links and respects access rights.

Access Vault

Stores passwords, keys and certificates. Access is role-based and requires approval; every request is logged, with rotation and revocation built in.

Dashboards

Management metrics from your accounting systems, with answers to questions about them in plain language. We connect to sources read-only.

Administration & Audit

Sign-in via AD/LDAP or SSO, roles, a log of user and agent actions, retention periods, load monitoring.

Agents without passwords

Agents work in your CRM, 1C and email — but they don’t know your passwords. They sign in to systems through the Vault and get only the permissions they need. The agent handles routine work on its own; where you decide, an action waits for an employee to confirm it. Every step is logged.

Architecture

How the platform is built

Four layers: channels, platform, models and deployment. The model can be replaced without reworking integrations; security and logging work on every layer.

Platform layers: channels; the platform with agents, knowledge, access vault and dashboards; replaceable models; deployment of your choice. Company systems on the left, security and audit across all layers on the right.
Channels, platform, models and deployment. Security and audit run across all layers.

You keep the keys

We build it turnkey — and hand over all the keys

  • Data

    On your servers or on dedicated capacity. Never sent to external AI services and never used to train third-party models.

  • Models

    Open models running on hardware dedicated to you. You can replace the model without reworking integrations.

  • Code and settings

    We hand over integrations, prompts, agent workflows and configurations together with documentation.

  • Access

    The Vault with your keys and passwords belongs to your company. Our support team gets access only with your approval, and every access is logged.

  • Exit plan

    Written into the contract: how to carry on without us if you decide to change vendors.

You choose where the platform runs

3–4× cheaper*

Flatstack infrastructure

Turnkey dedicated GPU capacity. No hardware to buy, quick start.

works offline

Your hardware

We size the configuration, install the platform on your servers and set it up.

flexible

Any other setup

Your cloud, a hybrid setup or a phased move from our infrastructure to your own hardware.

Privacy with verifiable terms

  1. Processing boundary. The model and data stay in the agreed infrastructure, with no transfer to external services.
  2. Access and retention. Corporate sign-in, document permissions, an activity log and agreed retention periods.
  3. Integration control. No public AI APIs and no training of third-party models on company data.

* Comparison of the GPU-hour price for a comparable GPU class with the public rates of major cloud providers. We’ll send a calculation for your configuration on request.

Where to start

Typical first use cases

Internal support

Answers on policies and instructions, with links to the current documents.

“How do I arrange a business trip?”

Working with documents

Drafts of letters and proposals, contract summaries, extraction of company details.

“Summarize the supply contract and highlight the penalties”

CRM and 1C assistant

Finding information in your systems and preparing tasks.

“Create follow-up tasks from the meeting”

Management analytics

Metrics from your accounting systems and answers to questions about them.

“Why did downtime increase in Q2?”

Case studies

Where AI is already at work

Reference solution

Call quality control

An AI agent checks every sales call against your company standard, quotes the conversation and gives one concrete recommendation. Recordings and speech recognition stay inside the perimeter.

  • 100% of calls instead of a sample
  • Scored against your sales standard
  • Call summary goes straight to the CRM

Implementation

From assessment to production rollout

  1. 01

    Assessment

    We pick the use cases to work on and check your data, security constraints and load. The result: a use-case map, architecture and budget.

  2. 02

    Pilot

    We deploy the environment and onboard a limited group. We measure answer quality and response time against criteria agreed before the start.

  3. 03

    Production rollout

    We run load tests, onboard all users and hand over the documentation.

  4. 04

    Training and support

    A course for employees, role-based workshops, a prompt library. Ongoing support and platform development.

Law and accountability

Legal requirements in the architecture, not in a slide deck

Russia — Federal Law 152-FZ on Personal Data

Personal data of Russian citizens is stored in Russia. Processing stays inside the agreed perimeter. Legal grounds, the parties’ roles and safeguards are agreed before the project starts.

Kazakhstan and Uzbekistan

Kazakhstan, Law No. 94-V: citizens’ personal data is kept on servers in Kazakhstan. Uzbekistan, Law ZRU-547: since 27 March 2026, mandatory localization applies to biometric, genetic and telecom-subscriber data.

AI regulation — Russian Federal Law 243-FZ

From 1 March 2027, AI system developers must meet requirements for security, operating rules and technical documentation. We build them in up front.

Critical-infrastructure, industry and certification requirements are assessed separately. Hosting in a given country does not by itself mean full compliance.

Economics

3–4×

Our own infrastructure is cheaper than major cloud providers

Flatstack infrastructure is our own know-how: dedicated GPU capacity costs 3–4× less than renting from major cloud providers*. We calculate implementation and support costs for your specific project.

Request an estimate

* Comparison of the GPU-hour price for a comparable GPU class with the public rates of major cloud providers. We’ll send a calculation for your configuration on request.

About us

20 years of building software. Now — AI that works inside your company

Flatstack is an engineering team from Kazan. Since 2006, we have built more than 100 websites and services and 50+ mobile apps. We work with banks, retailers, industrial companies, universities, football clubs and startups, in Russia and abroad. We use AI extensively in our own work: it shortens timelines and lowers project costs.

In the media

  • Реальное время
  • БИЗНЕС Online
  • Поток
  • SmartHead
  • Enter
  • DK.RU

FAQ

Frequently asked questions

Can the platform run without internet access?

Yes. On your hardware, the model, knowledge base and integrations run in an air-gapped perimeter, and we deliver updates offline.

Which models do you use?

We select an open model for your tasks, language requirements and available hardware. The model can be replaced without reworking integrations.

Do you train the model on our data?

No. The knowledge base grows with verified answers, but the model is not fine-tuned, and data is not shared with third parties.

Where will the platform run?

On your hardware, on our infrastructure or in any other agreed setup. You can start on our infrastructure and move to your own hardware later.

How much does it cost?

It depends on the number of users, the use cases and the deployment option. We’ll calculate it after a short call. Our infrastructure costs 3–4× less than capacity from major cloud providers.

How long does implementation take?

It depends on the number of departments and use cases. We’ll give you a rough estimate on the first call and fix the exact plan after the assessment.

Which systems does the platform work with?

Confluence and file storage, CRM, 1C, corporate email, telephony, AD/LDAP. Anything else we connect via API.

Who will see our data?

Only your employees, within their permissions. Our support team gets access under the contract, with your approval, and every access is logged.

Do you work outside Russia?

Yes — in Kazakhstan, Uzbekistan, Armenia and other countries, taking local data storage requirements into account.

First call: 30 minutes, online.

Let’s discuss your project

Tell us what you’d like to automate. We’ll reply within one business day and suggest a format: a demo, an assessment or a pilot.

We’ll sign an NDA before discussing any details.

Helps us store your request where your country’s law requires