AI AGENTCONCIERGE

SERVICE BOUNDARIES & DISCLOSURES

Clear expectations.
A better working relationship.

An audit, automation, integration or managed agent environment needs a clear agreement about what the agent can do, what it can access, what we maintain and what stays your responsibility. This page sets expectations; it is not a substitute for your client-specific agreement or a privacy notice.

Updated October 2026 · Applies to Audit & Roadmap, the AIC Executive Assistant, Custom Builds and Managed Care unless the client-specific agreement says otherwise.

01 / SCOPE & APPROVALS

Useful access.
Clear limits.

Discovery identifies possibilities; it does not grant permission. Before a workflow goes live, its tools, accounts, allowed actions and approval requirements are recorded and accepted.

  1. PROPOSE

    A useful workflow or improvement

  2. AGREE

    Scope, access, cost and review rules

  3. TEST

    Expected results and denied actions

  4. ENABLE

    Only the accepted configuration

Some actions may have standing authorization; others require approval each time. A draft is not permission to send it. A recommendation is not permission to change a live system. Payments, publication and changes to live systems need explicit rules in the build.

Telling an agent not to perform an action is not the same as technically preventing that action. Instructions and technical restrictions are different safeguards, and we describe the controls actually tested for your configuration.

An audit identifies options and priorities; it is not permission to change a system or a guarantee that implementation will produce a stated saving.

02 / DATA & ACCESS

Know where information goes.
Not one blanket setting.

Your agent works in a dedicated managed environment. External providers and connectors—the AI model provider, your email and calendar services, any approved third-party tools—each receive the information needed for the processing you approve. A separate environment is one part of the design; providers, connectors, backups, logs and maintenance access also matter.

YOUR APPROVED SOURCESBusiness informationOnly the systems and data in scope
YOUR AGENT’S WORKPLACEWork & useful contextDedicated managed environment
SELECTED PROVIDERSModels & connectorsInformation needed for approved processing
Illustrative data path. The actual destinations and access are documented for each client in the data and access record.
What does the data and access record cover?
  • Which information is kept in the managed environment and which information is sent to each selected provider or connector.
  • Who can access operational records, backups, conversations or secrets, including any approved support or emergency access.
  • Retention, correction, export and offboarding arrangements—including what deletion can and cannot remove from third-party systems.
  • Additional restrictions or specialist review needed for sensitive information.

These details are settled for your configuration. This page does not claim zero retention or operator-inaccessible conversations.

Does a business subscription settle privacy?

A suitable business plan can improve the provider-side privacy terms; for example, some providers state that business products are not used for model training by default. That commitment has a defined scope. It is not a guarantee about every connected service, retained record or support process. We review the selected plan, authentication method and current provider terms for the proposed build.

Each operational data type has an agreed source of truth. An automation does not silently replace the authoritative system.

03 / SENSITIVE INFORMATION

Some information needs
its own rules.

Credentials, payment details, customer records, regulated data and confidential business information are not handled by default. Where a workflow genuinely needs them, the handling, storage and provider path are agreed in writing first, with any specialist review the information requires. Do not send passwords, credentials or customer records through the website contact form.

04 / CONTINUITY & MEMORY

We assume sessions end.
We design so the work doesn’t.

Continuity means selected durable context, current work state, project status, approved resources and handoff information carry forward between supported sessions. It is not unlimited memory. Separate applications and interfaces may keep separate conversation histories; we do not promise that everything discussed in one place automatically appears everywhere else. Corrections to retained context are handled through the agreed process.

05 / MAINTENANCE & SECURITY RESPONSIBILITIES

What Managed Care covers.
What it doesn’t.

Managed Care covers the managed agent environment, according to scope: operating-system and runtime maintenance, patching, supported connector maintenance, troubleshooting, continuity-system maintenance, backup and recovery procedures, and vulnerability scanning of the managed agent environment. Vulnerability scanning means periodic checks of that environment for known weaknesses, with findings addressed within the agreed service; the frequency and remediation targets are set in your agreement.

Standard service is not complete cybersecurity management. It does not include antivirus or endpoint protection for your devices, a security operations centre, defence of your corporate network, identity security for your whole organization, incident response, or security for systems outside the managed agent environment. Those can be scoped separately if needed. Your own accounts, devices and credentials remain your responsibility.

06 / CLIENT & AIC ACCESS

Who can reach what.

You authorize the accounts and tools your agent may use, and you remain the account owner; some verification, consent or reauthorization steps may need you. AIC holds the access needed to build and maintain the managed environment. Any support or emergency access to operational records, backups or conversations is defined in the data and access record, not assumed. Access ends according to the offboarding terms in your agreement.

07 / COSTS & THIRD-PARTY SERVICES

Know what is included.
Know what needs a new decision.

Build

The Executive Assistant setup, or a Custom AIC build quoted to scope: discovery, configuration, agreed connections, testing and onboarding. Deliverables and acceptance criteria belong in the proposal.

Managed Care

The agreed environment maintenance, patching, connector upkeep, continuity maintenance, vulnerability scanning and troubleshooting. Support channels, hours and response targets are stated in your agreement—not assumed.

Third-party services

Hosting, your AI/model subscription and other applicable services are billed separately and identified up front. Usage limits, availability and authentication changes on those services can affect the workflow.

A suggestion is not a charge or permission to expand the build. New integrations, substantial development or changed service needs are reviewed together, with any scope or price change agreed before work proceeds. If AIC has to engineer something new for you, it is not included in the Executive Assistant.

Integration feasibility depends on provider access, API behavior, plan limits, licensing, authentication, and changes made by the provider. Audit & Roadmap is quoted to scope; timing and deliverables are agreed before work begins.

08 / SERVICE AVAILABILITY & RECOVERY

Designed to keep working.
Not guaranteed never to stop.

The managed environment is designed to stay available and to recover from ordinary failures, with backup and recovery procedures included in Managed Care. This page makes no uptime or response-time guarantee. Monitoring, recovery targets, support hours and response targets are set in your client-specific agreement, and third-party outages are outside our control.

09 / AGENT OUTPUT & HUMAN JUDGMENT

Review the result.

AI can make mistakes, omit details or misinterpret a request. Review important outputs and use qualified professional judgment where the work requires it. Greater responsibility is earned through testing and agreement; some tasks should always keep human review, and new access or actions require a new agreement.

AIC may recommend retaining a manual step where the process is unstable, the source data is unreliable, the task is consequential, or the expected return does not justify automation.

10 / EMERGING AND ADVANCED CAPABILITIES

Collaboration does not imply access.

Where a build uses more than one separately configured agent, each keeps its own working context, resources and permissions. Collaboration does not automatically give one agent another agent’s memory, resources or authority. Advanced agent-to-agent message-bus systems remain R&D / individually scoped and non-commercial by default. Controlled handoff work does not establish generalized production readiness.

Voice & Conversational Systems — scoped/emerging

Where voice fits the engagement, consent, recording and retention rules, identity disclosure, human escalation, transcription quality and provider dependence must be defined. Supported voice interaction in an assistant does not establish a generalized commercial Voice AI practice.

11 / EVIDENCE & CLAIMS

Keep the claims grounded.

Results published on this site are scoped to the workflow and environment they came from. Time saved is an estimate unless measured against a defined baseline, and we account for the owner’s review and correction time. Testing is specific to the environment being delivered and is not a general certification. See the current evidence.

12 / OWNERSHIP, HANDOFF & OFFBOARDING

Settled in writing.
Not assumed from a website.

Final ownership and offboarding terms are established in the client-specific agreement rather than assumed from this website. That agreement covers what you own, what is handed over, what is exported or removed at the end of the service, and any transition assistance. Our working direction is that you control the core accounts and infrastructure your agent uses; the specifics are confirmed for your build.

Before we start

The client-specific packet includes an accepted scope, a data and access record, commercial terms, support and recovery arrangements, and ownership, export and offboarding provisions. Unresolved choices remain open for agreement; the website does not settle them by implication.

Start the conversation