A buyer asks for evidence
You need to explain how your AI works, what controls are in place and what your documentation supports.
AI compliance audits by AI product engineers
Know what to fix before a buyer review or EU launch. We review your AI product’s technical controls and evidence, then give your engineers a prioritized action plan.
Start with your product URL and the reason for the review. No source code needed and no commitment to a paid audit.
When you need an audit
For teams with an AI product and a concrete compliance question, buyer review or launch requirement.
Your AI product is entering the EU
You need a technical readiness review against the EU AI Act requirements identified as relevant to your product.
Legal has identified requirements
Your engineering team needs to understand the missing controls, evidence and implementation work.
What you get from the audit
A product-specific technical review, written findings and a plan your team can implement. We agree which requirements, controls and evidence are in scope before work begins.
Architecture and data flows
A review of the AI features, model providers, integrations and data flows in scope, with assumptions to confirm with your legal team.
Controls and supporting evidence
Checks of the agreed controls, such as user disclosures, human approval, access permissions and sensitive-data handling. We record the evidence and what is missing.
Written technical report
A record of the scope, evidence reviewed, findings and limitations, so your product, engineering and legal teams can work from the same facts.
Prioritized engineering action plan
Practical remediation tasks and a walkthrough with your team. If you need help implementing changes, we can scope that work separately.
AI products we’ve built
Our development work includes LLM applications, semantic search and connected AI assistants. These are engineering projects, not compliance-audit case studies.
Semantic talent and company search
We built a platform with natural-language search, vector retrieval and LLM-assisted ranking, using FastAPI and PostgreSQL with pgvector.
This work involves the data flows, model outputs and application logic that technical reviews need to examine.
Read the search platform case studySakerCanine AI sales assistant
We built an e-commerce assistant that uses structured breed and product data to recommend harness sizes, with n8n workflows for lead and chat-data exports.
This work connects AI responses, customer inputs, business rules and downstream integrations.
Read the SakerCanine case studyWhat a finding looks like
Illustrative example, not a client result
An approval step exists on paper, but not in the product
Observation
The documented workflow requires approval before a CRM update, but the AI agent can write without it.
Evidence
The workflow specification requires approval. A test trace shows an update with no approval event recorded.
Recommended action
Enforce approval before writes, narrow the agent’s permissions and retest approved and rejected requests.
Start with one AI product or workflow
Agree a focused scope before committing to a larger review
Discuss your review
Share your product URL and what prompted the review. We clarify the task and propose a scope, fee and timeline. Paid work starts only after you agree.
Review your product
We assess the agreed technical controls and documentation, record findings and identify evidence that is missing.
Walk through the findings
You receive the report and prioritized action plan. We discuss them with your team. Implementation, if needed, is quoted separately.
Before we begin
No. This is a technical compliance-readiness audit. It does not certify compliance or replace legal advice. Final legal interpretations and applicability decisions should be confirmed with qualified counsel.
It is a paid engagement. We quote the scope, fee and timeline after reviewing your product and assessment needs. Sending a request does not commit you to a purchase.
Not to discuss the scope. Start with a public product or company URL and a short description. Any documentation or system access needed for the audit is agreed separately. Do not send credentials or sensitive customer data through this form.
Discuss your AI compliance audit
Tell us what your AI product does and what prompted the review. We’ll discuss what needs checking and agree the scope, fee and timeline before paid work begins.
Discuss your AI audit
Technical assessment, not legal advice or certification