Controlled AI automation for companies that need work to be clearer, safer, and easier to run.

Led by Róbert Kolesár
Founder & AI Automation Solution Architect

I build controlled AI automation systems that turn manual business work into reviewed, evidence-backed workflows.

Aureus helps companies turn scattered manual work into controlled workflows.

That means a real business process is mapped, repeated steps are automated, uncertain steps are routed to a person, and the system keeps a clear record of what happened.

AI prepares.
People approve.
The system keeps evidence.

The method behind this is Aureus OS. In plain language: it is a way of organizing AI, automation, GitHub, n8n, validation, approvals, and handoff so work does not become a pile of disconnected chat outputs and fragile workflows.

The private Aureus source-of-truth has been consolidated into Aureus OS: a single operating layer for mission intake, product/software proof work, n8n workflow governance, validation, evidence, preview boundaries, and approval-gated production promotion.

This public profile is the safe external version of that story. It shows what can be reviewed without exposing private workflows, credentials, endpoints, runtime logs, POHODA internals, client data, or production settings.

For the current public-safe status note, open Aureus OS final public status.

Safe CV wording is available in docs/portfolio/cv-usage.md.

Open the full menu: docs/portfolio/offer-menu.md

Example: a company has manual invoice or process work spread across email, spreadsheets, folders, and memory.

manual business chaos
-> process map
-> controlled AI workflow
-> human review boundary
-> proof notes
-> next-step proposal

Open the demo: docs/portfolio/public-demo-flow.md

For one concrete fictional example, open docs/portfolio/synthetic-demo-case.md.

Each proof path has its own visual because each system explains a different business problem.

Proof path	Visual	What it shows
Sales Machine		sales follow-up with qualification, approval, reply handling, and reporting
FinEcon		Pocket intake, document review, bridge handoff, proof notes, and accountant validation boundary
Aureus OS		how AI-assisted work becomes scoped, reviewed, validated, evidenced, and easy to hand off

Open the showroom: public-proof/README.md

The client-ready use-case system translates the portfolio into six sales and pilot conversations: Automation Audit, n8n Review + Build, FinEcon Pocket / Bridge, Sales Machine, Aureus OS / AOP, and Public Proof Website + Automation.

Start with the Use Case Offer Sheet, then open the English showcase copy, Slovak showcase copy, client-language PDF design spec, Instagram carousel export, or Git proof map.

Most automation projects stop at "the workflow runs."

Aureus cares about the part companies actually depend on:

• who owns the next step,
• what AI may prepare,
• what a person must approve,
• what happens when something fails,
• what evidence remains,
• and whether the team can understand the system after handoff.

That is the difference between a clever demo and a system a business can actually understand, review, and improve.

This profile shows architecture thinking, demo flows, proof-pack structure, case-study directions, offer boundaries, and review models.

It does not expose private implementation, real customer data, credentials, production context, or unsupported result claims. The full boundary is documented in docs/portfolio/public-boundary.md.

The root is intentionally clean. Supporting material lives in docs/portfolio and deeper proof packages live in public-proof.

This profile does not claim customer results, production deployments, accounting correctness, tax or legal advice, trading performance, official security certification, public access to private systems, or promised revenue / ROI.

It is a public portfolio for controlled AI automation design, review boundaries, proof-safe delivery, and business process clarity.

python scripts/validate-public-portfolio.py

Aureus Automation Lab turns manual business work into controlled AI-assisted systems with review, evidence, and clear next steps.