gayandesilva.com

What It Is

Gayan de Silva, PhD’s website presents an AI compliance audit and consulting service for the EU AI Act, rather than a general-purpose AI SaaS tool. Its core framework, ACAI, focuses on traditional ML, LLM systems, automated pipelines, and agentic workflows in production environments. The goal is to help companies identify compliance exposure, produce technical evidence, and build remediation plans before enforcement of high-risk AI requirements begins in August 2026.

Core Capabilities and Assessment Dimensions

The service stands out for emphasizing “real engineering testing” rather than purely document-based review. The traditional ML track includes AI system inventories, Annex III risk classification, ECE calibration testing, PSI distribution-drift detection, SHAP/LIME explanation faithfulness, and adversarial robustness testing. The LLM and Agentic AI track covers autonomy grading, structured logging, human override, prompt injection, tool-permission boundaries, and consistency under changing context. Each finding is mapped to specific EU AI Act articles and assigned a severity level.

Pricing and Deliverables

The website offers a free 30-minute discovery call, as well as a free 15-question diagnostic that does not require an email address. Paid services are offered in three tiers: Readiness Scan from €8,000, Full Compliance Audit from €25,000, and ongoing compliance advisory from €4,000/month. Deliverables include a findings register, a 90-day remediation roadmap, raw technical test results, a board-level compliance summary, and a 60-minute executive debrief. Overall, the pricing is more enterprise-oriented.

Pros and Cons

The strengths are a concrete methodology, granular technical testing, and coverage of emerging Agentic AI risks. It also states that the client team runs tests in its own environment, without requiring external access to production systems, which is more privacy- and security-friendly. The limitations are that the service depends heavily on cooperation from the client’s engineering team and is mainly centered on the EU AI Act, so it is not suitable as a general AI development platform or a unified global compliance solution. The website does not specify Chinese-language support, payment methods, or SLA details.

Who It’s For and Access from China

This is best suited to companies operating in the EU or providing AI systems to EU customers, especially in regulated sectors such as finance, manufacturing, energy, marketing technology, and supply chain. For Chinese companies bringing products into the European market, it could be considered as an external technical audit expert option for EU AI Act compliance. Direct access from China is unknown, and payment and contract arrangements are not disclosed. If a local alternative is needed, companies could consider major consulting/audit firms, model governance platforms, or an internal compliance team combined with open-source evaluation tools.