AI Exchange Trust Center
This page is about the platform, not about any capability published on it. It answers the questions a security, privacy or procurement review brings, in plain language; behind each answer, one step deeper, are the statements it rests on and how each one is known. None of it is a certification, and no certification is claimed.
AI Exchange operates the platform through which organizations describe a need, find and evaluate reusable AI capability, and interact under governed workflows. This page states the platform’s security and governance posture at the level a security, privacy or procurement review needs first. The questions below carry the current facts, each with how it is known; deeper evidence is supplied on request.
01
AI Exchange is responsible for the security and operation of the AI Exchange platform: its identity and authorization boundaries, its governed workflows, its evidence and publication controls, and the actions the platform performs. Responsibilities outside that platform boundary remain with the relevant participants unless AI Exchange explicitly assumes them under the applicable service or transaction terms.
02
Core data and model services are not directly exposed to the public Internet; the public boundary is served over HTTPS only. Workloads authenticate to platform infrastructure by identity rather than stored credentials. Each organization is an isolated authorization domain. Infrastructure is defined as code and released through a controlled process, and the deployed posture is checked against AI Exchange’s declared architecture and controls.
03
Platform data is processed and stored in the European Union on Microsoft Azure, with one documented exception for live-voice discovery. Individuals can obtain their data and record rights requests, and erasure runs through a governed, verified procedure. The subprocessors the platform relies on are stated.
04
The platform’s assistants analyse and recommend; they do not decide. Consequential actions — publishing a capability, selecting a proposal, contacting an external party — are authorized by a person. A search result, a fit assessment and a buyer’s approval are distinct records. Model access is governed through the platform’s own identity; application code holds no provider keys.
05
Publication is governed: a capability version reaches the marketplace only through a release gate the server enforces. Uploaded material is analysed in an isolated environment for known vulnerabilities, exposed secrets and insecure configuration before any assistant or reader may use it, and unresolved findings block publication. Evidence about a capability is recorded and shown with that capability, separately from this statement about the platform.
06
No independent certification, audit or penetration test is claimed today. The statements on this page rest on the platform’s own checks against the deployed service, observed behaviour and AI Exchange’s declarations, each marked with how it is known. A controlled, more detailed assurance record is supplied on request after a person has reviewed the request; nothing unlocks by signing in.
AI Exchange completed a self-assessment of the deployed Azure workload against the five pillars of the Microsoft Azure Well-Architected Framework on 2026-09-18.
The review found established controls in areas including identity-based access, private service boundaries, encryption at rest and in transit, infrastructure as code and controlled releases. It also identified material improvement priorities, particularly around service resilience and redundancy, recovery objectives and testing, and performance targets and capacity engineering.
Azure Advisor provides an additional ongoing configuration and operational signal for the deployed subscription.
This is an AI Exchange self-assessment, not a Microsoft review, certification or independent audit. Detailed scored results and remediation evidence are available through the deeper-assurance process.
Azure workload architecture
Well-Architected self-assessment across five pillars, 2026-09-18.
AI risk governance
Assessment against the NIST AI Risk Management Framework not yet performed.
Cloud control posture
Assessment against the CSA Cloud Controls Matrix not yet performed.
Material limitations a reviewer should weigh today
01
Declared by AI ExchangePublished documentNot claimed
AI Exchange operates the platform that connects business demand, reusable AI capability, evidence and governed interaction between participants. It is responsible for the security and operation of the AI Exchange platform, including its identity and authorization boundaries, governed workflows, evidence and publication controls, and actions performed by the platform. Responsibilities outside that platform boundary remain with the relevant participants unless AI Exchange explicitly assumes them under the applicable service or transaction terms. The operating legal entity, registration details and formal data-protection role remain subject to the published legal documentation and are identified there where still pending.
AI Exchange is responsible for the security and operation of the AI Exchange platform; responsibilities outside that platform boundary remain with the relevant participants unless AI Exchange explicitly assumes them under the applicable service or transaction terms.
As of 2026-09-15
Terms, Privacy, Cookies and Legal Notice are governed, versioned documents. English governs; no translation is approved. Each document states what is still pending before production.
As of 2026-09-15
Not claimed: The operating legal entity, its registration and its address.
Listed as pending before production in the Legal Notice.
Not claimed: The platform’s controller or processor role under GDPR, a data-protection contact, and contractual data-residency commitments.
Not yet decided. The processing locations above are the configured facts, not a commitment.
02
Declared by AI Exchange
Platform data is processed and stored in the European Union on Microsoft Azure: the application and its data in Azure Spain Central, the AI models in Azure Sweden Central. One exception is declared: the live voice discovery path is configured to allow processing outside the EU data zone, so EU-only processing is not guaranteed for that path. Database backups are kept in the same region for 7 days, without a geo-redundant copy.
The application and its data are deployed in Azure Spain Central; the AI models (Azure AI Foundry) in Azure Sweden Central. Both are EU regions.
As of 2026-09-15
Document understanding is bound to the EU data zone of the model resource. Live voice discovery is configured to allow global processing; EU-only processing is therefore not guaranteed for that path.
Where EU-only processing is required, live voice should not be used.
As of 2026-09-15
Database backups are retained for 7 days in the same region. Geo-redundant backup is off.
As of 2026-09-15
03
Declared by AI Exchange
Two: Microsoft Azure (hosting, database, storage and AI models) and Hostinger (transactional email such as verification and password reset). This is the set named by the platform’s own configuration. A reviewed subprocessor list with processing agreements is not yet published.
Hosting, database, storage and models: Microsoft Azure. Transactional email: Hostinger. No other processor is declared.
This is the set the configuration names, not a reviewed subprocessor list.
As of 2026-09-15
04
Observed on the live platformNot claimed
Accounts sign in with a verified email address and a password, and sessions are controlled by the server. Each organization is an isolated authorization domain: a private organization’s profile is withheld from other signed-in users, not only from anonymous visitors, and a session never confers authority over another organization. Password-reset requests reveal nothing about whether an address is registered. Single sign-on and multi-factor authentication are not offered today.
An account must verify its email address before it can sign in. Sign-in is refused until the address is verified.
As of 2026-09-15
A password-reset request for an unknown address is answered exactly like one for a known address, so the form cannot be used to discover accounts.
As of 2026-09-15
Organizations are separate tenants. A private organization’s profile is withheld from other signed-in users, not only from anonymous visitors: a session is not authorization over another organization.
As of 2026-09-14
Not claimed: Single sign-on or federated sign-in for customer accounts, and multi-factor authentication.
Customer sign-in is email and password with mandatory verification. No customer SSO or MFA is offered today.
05
Checked against the deployed platformDeclared by AI Exchange
The public boundary is served over HTTPS only. Core data and model services are not directly exposed to the public Internet. Workloads authenticate to platform infrastructure by identity rather than stored credentials, and storage and service access is identity-based. The infrastructure is defined as code, released through a controlled process, and its deployed posture is checked against AI Exchange’s declared architecture and controls.
HTTPS is enforced on the public surface.
As of 2026-09-15
Workloads authenticate to platform infrastructure by identity rather than stored credentials; secrets are held in a key store and referenced, never copied into configuration.
As of 2026-09-15
Infrastructure is defined as code, released through a controlled process, and its deployed posture is checked against AI Exchange’s declared architecture and controls.
This is the platform checking itself. It is not an independent audit.
As of 2026-09-15
06
Declared by AI ExchangeObserved on the live platform
Uploaded material is held in the platform’s EU storage and is not published unless the organization publishes a capability that uses it. Uploaded files and connected repository snapshots are analysed in an isolated environment for known vulnerabilities, exposed secrets and insecure configuration before any assistant or reader may use them; material that fails that analysis is never used. A capability version with unresolved findings cannot be published, and the refusal is enforced by the server. Antivirus-style malware scanning is not performed today.
Artifacts uploaded for a capability are assessed for known vulnerabilities in their components before publication, and the assessment is bound to the exact material assessed.
As of 2026-09-15
A capability version whose uploaded artifacts carry unresolved security findings cannot be published. The refusal is made by the server and cannot be overridden from the publishing form.
As of 2026-09-13
Uploaded files are analysed in an isolated environment for known vulnerabilities, exposed secrets and insecure configuration before any assistant or reader may use them; material that fails that analysis is never used. Antivirus-style malware scanning is not performed today.
As of 2026-09-18
07
Declared by AI Exchange
AI features run on model deployments that AI Exchange operates on Microsoft Azure AI Foundry, reached through a provider-neutral gateway. Model access is governed through the platform’s own identity; application code holds no provider keys, and no other model provider is used today. Prompts and documents are processed by those deployments in the EU, with the live-voice exception described under data location.
Model access is governed through a provider-neutral gateway to deployments AI Exchange operates on Azure AI Foundry, under the platform’s own identity; application code holds no provider keys, and no other model provider is used today.
As of 2026-09-15
08
Declared by AI ExchangeObserved on the live platformNot claimed
AI Exchange’s assistants help an organization describe a business need, help a builder describe a capability, and assess how well a capability fits a stated need. They produce assistance and recommendations, never binding decisions: a search result, a fit assessment and a buyer’s approval are distinct records, and fit is assessed against the organization’s own declared context. Publishing a capability version, selecting a proposal and contacting an external party on the platform’s behalf require a human decision; no automated component performs them. AI Exchange does not claim a formal model-evaluation programme today.
AI Exchange’s assistants help an organization describe a business need, help a builder describe a capability, and assess how well a capability fits a stated need. They produce assistance and recommendations, never binding decisions.
AI Exchange does not claim a formal model-evaluation programme today.
As of 2026-09-17
Discovery does not produce a verdict. A search result, a fit assessment and a buyer approval are distinct records, and fit is assessed against the organization’s own declared context and constraints.
As of 2026-09-15
Consequential actions require a human decision: publishing a capability version, selecting a proposal, and contacting an external party on the platform’s behalf. No automated component performs them.
As of 2026-09-15
Not claimed: The platform’s role under the EU AI Act, its transparency position for AI-generated output, and model evaluation results.
Not claimed: No assessment of the platform against the NIST AI Risk Management Framework has been performed yet.
The framework is held as a designated reference source; the assessment itself is pending.
09
Declared by AI ExchangeNot claimed
A signed-in person can download their personal data (access and portability) and record access, portability, erasure and rectification requests from their account page; each request carries a one-month deadline, and erasure is carried out by an authorised operator through a governed procedure whose result is verified before the request is closed. A person who cannot sign in uses the contact form. Retention periods have not yet been decided: data is kept while the account exists, nothing is expired automatically, and backups keep a copy for 7 days after erasure.
A signed-in person can download their personal data (access and portability) and record access, portability, erasure and rectification requests from their account page, each with a one-month deadline. Erasure is carried out by an authorised operator through a governed procedure whose result is verified before the request is closed.
A person who cannot sign in goes through the contact form. Backups keep a copy for 7 days after erasure.
As of 2026-09-17
Not claimed: Retention periods for account data, discovery conversations and uploaded material.
Not yet decided. Data is kept while the account exists and nothing is expired automatically until a period is decided.
10
Declared by AI ExchangeNot claimed
None today. AI Exchange does not claim SOC 2, ISO 27001 or any other certification, and claims no independent audit, penetration test or attestation. The statements on this page are self-assessed and evidence-backed, each marked with how it is known.
No certification — SOC 2, ISO 27001 or any other — is claimed, and no independent audit or attestation is claimed.
As of 2026-09-15
Not claimed: No assessment of the platform against the CSA Cloud Controls Matrix has been performed yet.
11
Not claimed
Security concerns can be submitted through the contact form, and are handled by the platform’s operators. A dedicated security contact, a published disclosure policy and a published incident-response process do not exist yet.
Not claimed: A dedicated security contact, a published vulnerability-disclosure policy, an incident-response process and penetration-test results.
To report a security concern today, use the contact form.
12
Declared by AI ExchangeNot claimedObserved on the live platform
Releases follow a controlled process: a release is prepared and verified by automation, and publishing it to a production environment is a decision reserved to a person. AI Exchange currently operates a single pre-launch environment; a separate production environment does not exist yet. No availability target, maintenance window or status page is published. A self-assessment of the deployed workload against the Azure Well-Architected Framework was completed on 2026-09-18; its scored results are supplied under deeper assurance.
AI Exchange currently operates a single pre-launch environment. No separate production environment exists yet.
As of 2026-09-15
Releases follow a controlled process: automation prepares and verifies a release; publishing it to a production environment is a decision reserved to a person.
As of 2026-09-15
Not claimed: Availability targets, maintenance windows and status reporting.
A self-assessment of the deployed workload against the five pillars of the Azure Well-Architected Framework was completed on 2026-09-18. It found established controls and material improvement priorities; the scored results are supplied under deeper assurance.
A self-assessment by AI Exchange, not a review by Microsoft or an independent party.
As of 2026-09-18
Azure Advisor provides an additional ongoing configuration and operational signal for the deployed subscription; its recommendations are tracked together with the review findings.
As of 2026-09-18
How to read this page
Confirmed by AI Exchange’s own automated checks against the deployed platform on the date shown. Not a certification.
2 statements
A behaviour observed on the deployed platform on the date shown.
6 statements
Stated by AI Exchange from its platform configuration, code or accepted design. Nothing outside the platform checked it.
16 statements
A versioned, human-approved document the platform publishes.
1 statement
Attested by an independent third party. None today.
0 statements
AI Exchange does not claim it. The answer says whether it is not offered, not decided, or pending.
9 statements
These are not the evidence classes used for capabilities — those are on each capability page and explained under principle 04. The vocabulary is different on purpose, so that "verified" means one thing on this site. The register behind this page is maintained by hand and re-read against its sources on the date shown.
Deeper assurance
A more detailed assurance record exists — the full statements behind each answer, their dates and boundaries, and a summary of the automated control sweep. It is supplied deliberately, to a named reviewer, after a person at AI Exchange has reviewed the request.
Signing in does not unlock it: a registered account is not a confidentiality agreement. A request is reviewed by a person and answered outside the product.
Tell us who is asking and why. A person at AI Exchange reviews every request; approved material is delivered outside the product, to you.
Capability evidence lives elsewhere
Each published capability shows what has been established about it by class — declared, verified, expert-reviewed, attested, observed, buyer-reported — with absence stated. A signed-in reader sees every record; an anonymous visitor sees the summary. Nothing on this page speaks for a capability.