Samo Jurdik Samo Jurdik

Kitchen design visualisations

AI phone call analysis

We used AI to convert initial kitchen CAD data into high-fidelity, lived-in design images.

Impact

  • The designer saves days of work in a professional rendering software.

  • He skips the hassle of organising professional photography sessions.

  • His clients have immediate feedback loop on how their home will look like.


CLIENT: Interior design studio

CLIENT REVENUE: 300k Euro

INDUSTRY: Construction

DURATION: 1 Week


Source image exported from drawing studio software

AI visualisation

Automatic generation of views from multiple angles

Or a short video clip

The Challenge

Contextual Gaps

AI needed to accurately infer and generate realistic environmental context, including lighting, textures, and domestic clutter, absent in the original file.

Maintaining Design Fidelity

The final photorealistic output had to strictly preserve the exact technical dimensions and core design elements from the initial draft.


The Solution

We used Gemini 3 Pro Image to understand the design file and then automatically create a realistic picture by adding details, textures, and context.

  • Developed a proprietary Gemini implementation, engineered for repeatable high-fidelity visualization and rapid, consistent output generation.

  • Reduced visualization time from days (manual) to minutes, achieving near real-time rendering of complex interior spaces.

  • Attained a high image realism score while maintaining 1:1 dimensional accuracy against the source CAD model.

Read More
Samo Jurdik Samo Jurdik

Phone call summaries by AI

AI phone call analysis

An AI phone call analysis system that transforms conversations into summaries, next steps, and recommendations for the operators.

Impact

No need for the company owner to listen to full length recorded phone calls. A quick look at the summaries is enough.

Actions overview and tracking for the phone call operators like to send an email to the customer, do a follow up call etc.

Sentiment analysis - was it a positive call? Is the customer happy?


CLIENT: Environmental consulting agency

CLIENT REVENUE: 0.5M Euro

INDUSTRY: Consulting

DURATION: 6 Weeks


The Challenge

Speech-to-Text Accuracy

Incorrect text input corrupts downstream AI tools (intent, solutions), leading to misunderstandings and inaccurate resolutions for the operator.

Peak call volumes cause the AI databases to time out

Operators lose real-time assistance.

Tone of Operator Recommendations

Technical, or demanding language causes operator confusion and decreased confidence.

The Solution

We built an automated n8n pipeline that leverages Google Speech‑to‑Text for accurate transcription and Google Gemini for real‑time call summaries, action items, and operator‑friendly recommendations.

  • Google outperformed ElevenLabs in our tests.

  • In order to avoid timeouts of the AI tools, we had to separate the tasks into multiple steps. One-by-one execution helped to mitigate the problem.

  • The tone of languages of the AI recommendations to the phone call operators proved to be delicate and it required extra attention.

Read More
Samo Jurdik Samo Jurdik

Supply-Demand Tool by AI

It all begins with an idea.

A predictive supply-demand tool that aligns procurement with real-world sales, inventory, and lead times.


CLIENT: Electronics manufacturer

INDUSTRY: EV Mobility

DURATION: 3 Weeks


The Challenge

Variant-Driven Complexity

The client produces multiple EV charger models, each selling at different proportions. Forecasting demand wasn’t just about total volume—it required understanding how variant mix affects component needs.

Fragmented Inventory Visibility

Inventory was spread across several categories: finished goods in stock, products ordered but not yet delivered, and components already ordered but still in transit. This made it difficult to assess true availability and plan ahead.

Lead Time and Procurement Uncertainty

Each component came with its own supplier lead time, and delays could ripple through the production schedule. The key question was: which components should be ordered, when, and in what quantities to meet future demand without overstocking or stalling production?

The Solution

We built a custom supply-demand planning engine that integrates real-time sales data, inventory status, and supplier lead times to generate precise procurement recommendations.

  • Product variant sales proportions

    Existing stock of finished goods and components

    Ordered but undelivered products and parts

    Component-specific lead times and supplier variability

    Minimum order quantities and batch sizes

  • Probabilistic demand modeling to simulate variant-level sales

    Constraint-based optimization for procurement timing and quantities

    Rule-based logic for deterministic checks (e.g., safety stock thresholds)

    Scenario simulation to test different sales and supply conditions

  • A modular interface for inputting assumptions and constraints

    Visual outputs showing component needs over time

    Editable forecasts and override options for planners

    Exportable recommendations for procurement systems

Operational Impact

90% less time in Excel

100% accuracy of the calculation

Business Impact

750K Euro annual cost savings due to inventory optimisation

24K Euro annual cost savings due to time saved in Excel

Read More
Samo Jurdik Samo Jurdik

Tech Support Powered by AI

It all begins with an idea.

A support assistant that reduced response times by 87% using only verified documentation.


CLIENT: Smart home system manufacturer

INDUSTRY: IoT & Home Automation

DURATION: 8 Weeks


The Challenge

Documentation Gaps

Documentation was fragmented across multiple product generations and sources. Support teams couldn't find accurate information quickly.

Hallucination Risk

AI models generate incorrect answers when lacking information. In technical support, wrong instructions could damage devices or create safety hazards.

The Solution

We implemented a RAG system that grounds every response in verified documentation.

  • Vectorised all documentation and manuals into a searchable semantic database.

  • Vector similarity search retrieves relevant documentation for each query.

  • AI answers only from provided context, with clear handling when information is unavailable.

Operational Impact

24/7 availability with instant responses

92% Customer Satisfaction

Business Impact

234K Euro annual cost savings due to reduced support staff overhead

Read More