We put AI to work, not on display.

AI or smart analytics features are at work in every one of our products: from a sales assistant to document reading, from audit reports to consumption forecasting.

We use AI at three levels.

  • We build with AI

    Our products are built through an AI-assisted engineering process. AI speeds up our team; our engineers always make the final decision.

    • Code analysis and review
    • Automated test generation
    • Documentation and architecture records
    • Debugging and root cause analysis
  • We build it into our products

    From sales to document management, AI and smart analytics in our products work to make users' jobs tangibly easier.

    • AI or smart analytics in 5 of our 5 products
    • AI models such as language models and OCR in 3 products
    • 5 AI and 16 smart analytics features
    List of features
  • We integrate it into your organization

    We add AI capabilities to your existing systems and build solutions that handle your data securely and deliver measurable benefits.

    • Document reading and information extraction
    • In-house knowledge assistant
    • Forecasting and risk scoring
    • Smart matching and classification

Smart features that work today.

Every feature below actually exists in our products. We have clearly marked which ones are AI and which are statistical or rule-based.

AI
Features that run on a real AI model, such as a language model or OCR.
Smart analytics
Features that run on statistics, forecasting and a rules engine.

Store Management

Home appliance and electronics retail

AI features: 3, smart analytics features: 3

Store Management page
  • AI

    Bundle assistant. Describe the customer's needs in free text; AI extracts the budget and constraints, and the system builds Economy, Balanced and Premium bundle options.

  • AI

    Natural language reports. Routes a question like "What did we sell most this month?" to the right report. No free-form queries are run, so data stays safe.

  • AI

    Sales arguments and promise tracking. When comparing products, it generates the differences to explain to the customer; it extracts promised dates and amounts from meeting notes.

  • Smart analytics

    Demand forecasting. The right stock at the right time, with statistical demand forecasting that also accounts for the wedding season.

  • Smart analytics

    Anomaly alerts. Unusual discounts, frequent cancellations and returns, cash discrepancy patterns and after-hours price changes are caught automatically.

  • Smart analytics

    Collection priority. The "Call today" list is sorted by delay and risk, so the team spends its time on the right customers.

Language model features are optional and must be turned on. Personal data is masked before processing, and every AI output is marked as a "Suggestion" and submitted for human approval.

Station Platform

Fuel stations

Smart analytics features: 4

Station Platform page
  • Smart analytics

    Consumption forecast and order suggestions. A consumption profile built for every day of the week and every hour predicts when the tank will run out and suggests an order.

  • Smart analytics

    Reconciliation risk classification. Wet stock reconciliation is automatically classified as Normal, Watch, Suspicious or Critical.

  • Smart analytics

    Meter reading anomaly alert. Discrepancies between meter readings and sales trigger an alarm the moment they occur.

  • Smart analytics

    Learning invoice matching. Incoming invoice lines are assigned to the right accounts with suggestions learned from past usage.

Foundation Management System

Foundations and civil society

AI features: 1, smart analytics features: 2

Foundation Management System page
  • AI

    AI-based OCR. Scanned documents are converted to Turkish text with neural network-based OCR; every document in the archive can be searched in seconds.

  • Smart analytics

    Duplicate record detection. Duplicate records of the same person are caught by checking name similarity, date of birth and phone number together.

  • Smart analytics

    Scholarship scoring. Applications are scored automatically against the criteria set by the foundation; the committee works from a ready-made ranking when deciding.

KiraTakip

Real estate and rentals

Smart analytics features: 3

KiraTakip page
  • Smart analytics

    Smart bank statement matching. Each line of an Excel or CSV statement gets a tenant suggestion scored by IBAN, tax number and amount; suggestions are approved in bulk.

  • Smart analytics

    Cash flow forecast. A conservative cash flow forecast based on the actual collection rate of the last six months.

  • Smart analytics

    Payment habit score. Tenants' payment behavior is scored; risky tenants are flagged before a payment is late.

Elfi ERP

Construction and contracting

AI features: 1, smart analytics features: 4

Elfi ERP page
  • AI

    Management analysis report. Findings from the audit checks are compared with the previous day; the connected language model (Claude, GPT or Gemini) explains them to management in plain language.

  • Smart analytics

    Payment integrity check. Suspected duplicate payments of the same amount to the same payee, requests split up within a short period and payments with missing documents are flagged.

  • Smart analytics

    Procurement check. Repeat purchases with a marked price increase, purchases made with fewer than three quotes and cases where the lowest quote was not chosen are highlighted.

  • Smart analytics

    Site alert center. Fuel consumption above machine norms, critical tank and silo levels, equipment due for maintenance and cafeteria counts that don't match timesheets, all in a single list; thresholds are adjustable.

  • Smart analytics

    Report consistency. Daily site reports are compared with timesheet and fuel records; reports that copy one another are caught.

AI commentary is optional: it is activated when the organization connects its own model key, which is stored encrypted. Only rule-based metrics are shown in the live demo.

We are aware of the responsibility that comes with power.

The principles we follow when bringing AI together with organizations' sensitive data.

Humans have the final say
AI outputs are marked as "suggestions" and submitted for user approval. Critical decisions are made by an authorized person, not by the system.
Privacy first
Features that connect to external AI services are off by default and work only if you turn them on. We state clearly on each product's page which data leaves the system. KVKK is taken into account from the first day of design.
Clear limits
AI does not run free-form queries on the database; it only triggers defined operations whose permissions have been checked.
Transparent
We state clearly what is AI and what is statistical or rule-based. We present our products through features that work, not through exaggerated promises.
Provider-independent
Our architecture is designed to work with different language model providers, so you are not locked into a single company.
Optional
Language model features are turned on at your organization's choice. The system also works fully with them turned off.

We publish summaries of selected news from the world of AI in Turkish only: AI news (in Turkish).

What can AI do for your organization?

Let's review your processes together and lay out concretely where AI will really make a difference and what it will cost.