The Multi-Agent Architecture of Stroncami.it and the Transition from UX to AX

The share of web traffic generated by bots has surpassed that by humans (over 51%). By the end of 2026, it is estimated that 40% of enterprise applications will integrate task-specific AI agents. In this scenario, traditional User Experience (UX) metrics become insufficient for Large Language Models. This is where Agent Experience (AX) comes in: the discipline that designs optimized digital environments so that autonomous agents can interpret them, trust the data, and act efficiently.

Stroncami.it, developed by Vicedomini Softworks, is one of the first platforms in the world to embody this new paradigm. It's not just a validator: it's a true multi-step LLM agent capable of performing comprehensive audits, measuring accessibility for both humans and artificial intelligences, and delivering a merciless but valuable technical “Roast”.

1. From UX to AX: The New Web Paradigm

While UX focuses on visual hierarchies, colors, and human interactions, AI agents primarily consume the structural DOM. Agent Experience (AX) positions itself between UX and Developer Experience (DX), prioritizing semantic purity, predictability, and actionability.

Discover our software consulting services → Custom Software Consulting and Development

2. How Stroncami.it Works

After rigorous validation (HTTPS + execution in a real browser with JavaScript), the system distributes the analysis across five main categories:

Try Stroncami.it now → stroncami.it

3. Multi-Agent Architecture

Stroncami.it uses an autonomous cycle of Planning → Tool Use → Reflection → Orchestration, overcoming the limitations of traditional prompts and simple RAG. This architecture allows it to handle DOM ambiguity, dark patterns, and dynamic changes with high resilience.

4. MCP and Agent Skills

The platform verifies both support for Anthropic's Model Context Protocol (MCP) and the use of Agent Skills (knowledge packets in Markdown/YAML), the two pillars for effectively connecting LLMs to the external world.

5. Human and Agentic Accessibility: the same path

Semantic HTML, correct ARIA attributes, and structured Schema.org data (JSON-LD) simultaneously improve the experience for screen readers and AI agents.

6. llms.txt: the business card for agents

Stroncami.it also evaluates the presence and quality of /llms.txt and /llms-full.txt, the emerging standard that is becoming essential for interoperability with intelligent agents.

7. The AX Score (from 1 to 10)

The final audit result is the AX Score, a rating from 1 to 10 (like in school):

• 1-4: Agent-Hostile (strong resistance to automation)
• 5-7: Needs Improvement (works but with inefficiencies)
• 8-10: Agent-Ready (excellent readiness for the agentic era)

Conclusion

Optimizing a site for Agent Experience is no longer an option: it's a strategy for survival and competitive advantage in the Agentic Web. Domains that ignore semantics, structured metadata, and discovery protocols risk becoming invisible to AI agents.

Vicedomini Softworks accompanies you in this transformation with deep technical audits, AX-oriented refactoring, and sustainable architectures without vendor lock-in.

Ready to roast (and improve) your site?
→ Start an audit with Stroncami.it
→ Request technical consulting

Acknowledgements
Special thanks to Regolo.ai for the powerful models used behind the scenes, particularly Brick-v1-Beta, which made Stroncami.it's multi-agent architecture possible.