For overarching managed services terms (support model, response targets, limitations, and client responsibilities), please refer to the Managed Services Service Description. A Glossary of Common AI Terms can be found at the end of this document.

Managed AI Program Overview

The Managed AI Program is designed to help clients manage AI use, reduce AI risk, support adoption, and assist in maintaining deployed AI infrastructure. The Managed AI Program provides clients with a structured operating model for more secure, effective, and responsible AI adoption.

The service is delivered through one of three program levels, selected based on the client’s AI maturity and desired capabilities: Secure & Enabled, Productive AI, or Intelligent AI. The table below summarizes the services included at each level.

Managed AI Program	Secure & Enabled	Productive AI	Intelligent AI
Safe & Secure Use
AI Workspace Guidance	Included	Included	Included
AI Acceptable Use Policy	Included	Included	Included
Shadow AI Management, monthly report	Included	Included	Included
Suspect Shadow AI Alerting	–	–	Included
Monthly Employee AI Prompt Reporting	–	–	Included
Productivity & Adoption
AI Maturity Score Assessment – Self Guided	Included	Included	Included
AI Business Analyst engagement	Community Engagement	Quarterly AI Alignment	Monthly AI Coaching & Quarterly AI Alignment
ITS Managed AI Hub	Included	Included	Included
Monthly AI Tip Drip	Included	Included	Included
Monthly AI Community Open Office Hours	Included	Included	Included
AI Video Learning Library	Included	Included	Included
Solutions Catalog	Includes 3 solutions	Includes 5 solutions	Includes all solutions
Usage and adoption reporting	Quarterly Report	Monthly Report	Monthly Review
Internal Champion Training	–	–	Includes 1 Training
AI Infrastructure Management
Managed AI Support including agents, AI workspace, connections, and integrations	Included	Included	Included
Managed MCP Server Connections	–	Includes 3 connections	Includes 5 connections
Non-MCP System Assisted Connections	–	Includes 2 connections	Includes 5 connections
Managed Private LLM	–	–	Included
Managed Retrieval Augmented Server	–	–	Included

Note: Hosted infrastructure, custom solutions, off-catalog connectors, and specialized deployments may require separate scoping, additional infrastructure, or a Professional Services engagement as described in the sections below.

Onboarding and Orientation

Managed AI onboarding prepares the client’s approved users, AI Workspace, policies, selected solutions, and support path for go-live. Based on the client’s selected Managed AI Program level, ITS coordinates discovery, configuration, orientation, activation resources, validation, and required handoffs.

How it Works

Onboarding begins with confirmation of the selected Managed AI Program level, user count, client point of contact, client-selected AI Workspace, business goals, approved use cases, security requirements, and any required SSO, connection, or infrastructure decisions. ITS then configures the Managed AI environment, provisions approved users, applies baseline governance controls, and deploys the selected catalog solutions and supported connections within the client’s plan limits.

Client orientation is delivered as a recorded live session, up to one hour, covering login, SSO or MFA, the client-approved AI Workspace, the approved AI Acceptable Use Policy, selected solutions, connected tools, and available enablement resources. After validation, ITS confirms production readiness, documents the managed AI environment, and transitions managed components to the service desk support path.

Frequently Asked Questions

What onboarding activities are included?
Included onboarding activities are limited to the client’s selected Managed AI Program and approved scope. This includes discovery, configuration of the managed AI environment, user provisioning, baseline governance setup, selected catalog solution deployment, orientation, validation, documentation, and service desk handoff.

What onboarding requests may require a separate project?
Requests outside the selected program or approved onboarding scope may require separate scoping and billing. Examples include custom AI agents or workflows, off-catalog connectors, additional connections beyond plan limits, hosted MCP, Private LLM, or RAG infrastructure deployment, manual RAG document curation, or complex client-specific integration requirements.

What client responsibilities are not included in onboarding?
The client remains responsible for internal approvals, legal or compliance review, policy acceptance recordkeeping, identifying users and business owners, validating source content, confirming access permissions, and ensuring AI use aligns with applicable business, legal, regulatory, and industry requirements.

[return to top]

Safe and Secure AI Use

The Safe and Secure AI Use component helps clients determine the ideal AI Workspace, define baseline acceptable use expectations, and gain visibility into AI tools being used across the organization.

The service is designed to help reduce unmanaged AI activity and support more responsible AI use, but it does not replace the client’s internal governance, legal, compliance, or data management responsibilities.

How it Works

All three stages include baseline Safe and Secure AI Use capabilities. The Intelligent AI stage adds additional Shadow AI visibility and reporting as noted below.

AI Workspace Guidance. The first step in deploying the program is for the client to designate its approved AI Workspace with the AI Business Analysts guidance. For purposes of this service, the AI Workspace is the company-managed environment employees use for day-to-day AI activity, such as ITS AI Workspace, M365 Copilot, Anthropic Claude/Cowork, or OpenAI ChatGPT. ITS supports the client-approved AI Workspace within the selected program scope. ITS AI Workspace is available for an additional monthly charge.

Acceptable Use Policy support. ITS provides a baseline AI Acceptable Use Policy template for the client’s leadership and legal team to review. The client remains responsible for review, approval, policy acceptance recordkeeping, and retention.

Shadow AI visibility and reporting. ITS inventories AI tools used across managed devices and provides reporting to help identify sanctioned and unsanctioned AI activity. (Please note: Only the Intelligent AI stage adds blocked-site review, suspected Shadow AI alerting, and monthly User Prompt Insight reporting.)

Frequently Asked Questions

Does this service guarantee that employees will not use unauthorized AI tools?
The service provides governance controls, visibility, reporting, and stage-based alerting to help reduce unmanaged or unauthorized AI use. It does not guarantee that all unauthorized AI activity will be prevented or detected.

Who approves the AI Acceptable Use Policy?
ITS provides a baseline AI Acceptable Use Policy template and can deploy the approved policy through the AI governance platform. The client remains responsible for leadership/legal review, approval, policy acceptance recordkeeping, and retention.

What Shadow AI reporting is included?
All Managed AI stages include monthly Shadow AI reporting that identifies AI tools detected across managed devices. Intelligent AI adds blocked-site review and suspected Shadow AI alerting between reports.

[return to top]

Productivity and Adoption

Productivity and Adoption services provide enablement, usage visibility, and guided resources to help clients increase responsible AI use over time. These services are intended to support adoption throughout the client’s organization.

How it Works

The Managed AI Program includes adoption and enablement resources designed to support ongoing AI use after deployment. Entitlements vary by program level, as described below.

• Secure & Enabled clients receive self-service AI maturity scoring, monthly tips, open office hours, ITS AI video library access, three catalogued solutions, and a quarterly automated usage and adoption snapshot.
• Productive AI clients receive self-service AI maturity scoring, monthly tips, open office hours, ITS AI video library access, up to five catalogued solutions, monthly usage and adoption reporting, and quarterly AI Business Analyst strategy access.
• Intelligent AI clients receive self-service AI maturity scoring plus a formal quarterly re-score reviewed with the AI Business Analyst, monthly tips, open office hours, access to the ITS AI video library, all current and future catalog solutions, monthly usage and adoption reporting reviewed with the AI Business Analyst, one internal champion training session, and monthly proactive coaching.

The ITS Solutions Catalog contains pre-built AI solutions organized by industry, role, and function. Before deployment, ITS validates each solution against the client’s use case and readiness to support alignment with business needs and a repeatable delivery model. Catalog access does not include custom solution design, off-catalog integrations, or deployments beyond the limits of the client’s selected program level unless separately scoped.

Frequently Asked Questions

Is AI maturity scoring included?
Yes. Managed clients can complete the AI Maturity Score Assessment through the ITS website as a self-guided assessment. Intelligent AI clients also receive a formal quarterly re-score reviewed with their AI Business Analyst.

Are open office hours included for all client stages?
Yes. Every managed client’s staff can join monthly live Teams open office hours hosted by an AI Business Analyst. Sessions are group-format and recorded for review in the Managed AI Hub.

Does the Managed AI Program guarantee adoption or productivity improvements?
The program provides enablement resources, reporting, and stage-based advisory support to help clients improve AI usage over time, but the client remains responsible for internal adoption, change management, and business outcomes.

Are custom solutions included in Productivity and Adoption?
No. Productivity and Adoption includes access to catalog solutions based on the client’s selected program stage. Custom solutions, off-catalog workflows, and expanded deployments require separate scoping.

How much access does a client have to the AI Business Analyst?

Access depends on the client’s selected program level and stage of its AI journey. The AI Business Analyst is available to all clients during open office hours. Productive AI and Intelligent AI clients also receive about 60 minutes and 90 minutes, respectively, of proactive attention on average each quarter.

How does the AI Business Analyst work with the Strategic Advisor, Technical Alignment Engineer (TAE), and vCISO?
The AI Business Analyst complements these roles by focusing on AI adoption, responsible use, usage insights, and AI solution planning. Strategic Advisors provide broader technology direction, TAEs identify technical alignment needs, and vCISOs advise on cybersecurity strategy and risk. The AI Business Analyst helps connect these inputs to AI-specific adoption and planning activities based on the client’s subscribed stage.

[return to top]

AI Infrastructure Management

AI Infrastructure Management covers support for the managed AI components deployed or configured by ITS within the client’s selected stage or separately scoped engagement.

What Makes Up an AI Component?

• Approved AI Workspace
• Deployed catalog solutions
• Approved connections
• Assisted integrations
• Eligible AI infrastructure after deployment

Examples of What Is Not Included

• Purchase of hosted infrastructure
• Custom solution development
• Off-catalog connectors
• Additional connections or expansion beyond plan limits
• Data warehouse or data lake
• Any item not included in the selected program stage

How it Works

ITS supports deployed managed AI components, so they remain operational within the limits of the client’s selected program level and the capabilities of the underlying platforms. Included support varies by program level, as outlined below:

• Secure & Enabled includes support for the client-approved AI Workspace, three deployed catalog solutions, and other managed AI components included with the program.
• Productive AI includes support for the client-approved AI Workspace, up to five catalog solutions, up to three managed MCP connections where required hosted infrastructure has been purchased, and up to two assisted app connections.
• Intelligent AI includes support for the client-approved AI Workspace, all current and future catalog solutions, up to five managed MCP connections, up to five assisted app connections, and managed Private LLM or RAG services where hosted infrastructure and deployment have been separately purchased or scoped.

Note: Custom solutions, off-catalog connectors, connections beyond plan limits, hosted infrastructure purchases, Private LLM deployments, RAG deployments, and specialized client-specific integrations require separate scoping and a Professional Services engagement. After a custom or separately scoped component is deployed, your managed service will provide ongoing managed infrastructure support only if the component is expressly included in the approved Managed AI Program support scope.

Frequently Asked Questions

What does managed AI infrastructure support cover?
ITS supports managed AI components it deploys or configures, including workspace components, catalog solutions, approved connections, and eligible integrations. Support is limited to the client’s selected program and approved Managed AI Program support scope.

Does the service desk support include prompt coaching or usage questions?
Prompt coaching and usage questions are handled through monthly open office hours, not the service desk.

Is the hosting of the MCP, Private LLM, or RAG servers included in the monthly program fee?
Required hosted infrastructure and related costs must be purchased separately. Deployment of Private LLM and RAG services requires separate scoping, and implementation services may be billed separately.

Out of Scope / Not Included

The items below are common examples of services, responsibilities, and outcomes that are not included in the Managed AI Program unless they are expressly included in the client’s selected program level or separately scoped and approved. This list is not intended to be comprehensive. If a feature or activity is not identified as included in the applicable program level or approved scope, it is not included in the service.

• Client governance and legal responsibilities: Client legal, regulatory, compliance, or industry-specific review; approval of AI policies, tools, use cases, access, and permissions; policy acceptance recordkeeping and retention; and ownership of AI strategy and governance decisions.
• Guaranteed outcomes: Guaranteed productivity gains, business outcomes, adoption levels, prevention or detection of all unauthorized AI activity, or accuracy of AI-generated outputs.
• Client data and decision responsibility: Data integrity, source-content quality, document or knowledge-source management, validation of client-provided data or AI-generated outputs, and responsibility for business decisions made using those outputs.
• Custom or expanded solution work: Custom AI solutions, off-catalog workflows or connectors, additional connections or deployments beyond program limits, complex client-specific integrations, manual RAG document curation, and new agents, workflows, or enhancements unless separately scoped.
• Hosted infrastructure and implementation: Purchase or ongoing client cloud charges for required MCP, Private LLM, or RAG infrastructure, and related implementation or specialized deployment services unless separately scoped.
• Training, adoption, and change management: Department-specific training, role-based workshops, adoption campaigns, organizational change management, or prompt coaching through the service desk. Prompt coaching and usage questions are handled through open office hours.
• External or unmanaged tools: Governance, management, and support of external, unmanaged, or non-ITS AI tools are excluded unless expressly included in a separately scoped and approved engagement.

[return to top]

Support

How it Works

All Managed AI Program plans include support through ITS’s existing service desk and standard Response Time Expectations. Support is limited to managed AI infrastructure issues, workspace access issues, deployed catalogued solutions, approved connections, and other components included in the client’s selected program or approved Managed AI Program support scope.

Support Boundaries

The service desk provides intake, triage, escalation, and resolution support for managed AI components deployed or configured by ITS. Support does not include prompt coaching, AI usage questions, custom solution design, unmanaged third-party AI tools, data integrity or source-content validation, business process redesign, adoption campaigns, or change management services unless separately scoped.

Frequently Asked Questions

What should be submitted to the service desk?
Service desk tickets should be used for workspace access issues, deployed solution failures, connection issues, integration errors, or other operational issues with managed AI components included in the client’s selected stage or approved Managed AI Program support scope.

Where do users go for prompt help or AI usage questions?
Prompt help and AI usage questions are handled through monthly open office hours, not the service desk.

Are requests for new agents, workflows, or enhancements handled through support?
The service desk may intake and route requests for new agents, workflows, or enhancements, but design, build, deployment, or expansion work outside the approved scope requires separate scoping and may require a Professional Services engagement.

Does support include client data validation or AI output review?
The client remains responsible for data integrity, source-content management, validating AI-generated outputs, and decisions made using AI-generated outputs.

Does ITS support AI tools that are not part of the Managed AI Program?
External, unmanaged, or non-ITS AI tools are not supported unless specifically included in a separately scoped and approved engagement.

[return to top]

Client Responsibilities

The Managed Services Service Description continues to define standard ITS support and general client responsibilities. The client responsibilities below identify items that are specific to the Managed AI Program and different from, or additive to, the standard managed services terms.

• Maintain fully managed or co-managed ITS services, including ITS-managed endpoints, infrastructure, cybersecurity, and service desk support, as required for delivery of Managed AI governance, reporting, and ongoing support.
• Approve the client-approved AI Workspace, AI Acceptable Use Policy, users, business owners, use cases, access permissions, and required source systems.
• Own AI strategy, governance decisions, adoption management, organizational change management, and business outcomes.
• Maintain data integrity, source-content quality, permissions, exclusions, and knowledge-source ownership.
• Validate AI-generated outputs before use and remain responsible for business decisions made using AI-generated outputs.
• Ensure AI use complies with applicable legal, regulatory, contractual, security, privacy, and industry-specific requirements.
• Assume responsibility for external, unmanaged, or non-ITS AI tools unless their governance, management, or support is included in a separately scoped and approved engagement.

 

 

Glossary of Acronyms and AI-Specific Terms

The following glossary defines acronyms and AI-specific terms used in this Managed AI Program service description. Definitions are written for client-facing clarity and are intended to support consistent interpretation of the service scope.

Term	Definition
AI	Artificial Intelligence; technology that can assist with tasks such as content generation, analysis, classification, summarization, recommendations, automation, and decision support.
AI Acceptable Use Policy	A client-approved policy that defines acceptable and prohibited AI use, including expectations for data handling, user behavior, security, privacy, and responsible use.
AI Agent	An AI-enabled system designed to pursue a goal, use tools or data sources, and complete defined tasks or workflows with varying levels of human oversight.
AI Business Analyst	An ITS role that provides AI-focused enablement, usage insight review, solution planning, and stage-based advisory support for Managed AI Program clients.
AI Governance	The policies, controls, responsibilities, reporting, and oversight used to guide safe, secure, compliant, and responsible AI use within an organization.
AI Maturity Score Assessment	A structured assessment used to evaluate an organization’s AI readiness, adoption maturity, governance practices, and opportunities for improvement.
AI Workspace	The approved environment employees use for day-to-day AI activity, such as ITS AI Workspace or another client-selected AI Workspace such as M365 Copilot, Anthropic Claude/Cowork, or OpenAI ChatGPT.
API	Application Programming Interface; a defined way for software applications or services to connect, exchange information, and trigger actions.
Assisted App Connection	A supported connection that allows an AI workspace or managed AI component to access or interact with an approved business application, subject to available permissions and platform capabilities.
AWS	Amazon Web Services; a cloud platform that may be used for hosted infrastructure when separately scoped and approved.
Azure	Microsoft Azure; a cloud platform that may be used for hosted infrastructure when separately scoped and approved.
Catalog Solution	A pre-built AI solution offered through the ITS Solutions Catalog and deployed according to the client’s selected stage, readiness, and approved use case.
Context Window	The maximum amount of information an AI model can consider at one time when generating a response.
Custom Solution	An AI agent, workflow, connector, or automation that is not part of the standard ITS Solutions Catalog and requires separate scoping, design, and pricing.
Fine-Tuning	Further training an AI model on specialized examples or data to improve performance for a particular task, domain, or style.
Guardrails	Rules, controls, limits, or safety mechanisms intended to reduce unwanted AI behavior, restrict access, and guide appropriate outputs or actions.
Hallucination	An AI-generated response that appears plausible but is incorrect, unsupported, incomplete, or fabricated.
Hosted Infrastructure	Cloud or server resources needed to run certain AI services, such as MCP servers, Private LLM environments, or RAG servers. These resources are not included in the monthly Managed AI Program fee unless separately scoped.
Human-in-the-Loop	A workflow design where a person reviews, approves, or validates an AI action or output before it proceeds.
Integration	A configured connection between systems that allows approved data access, task execution, or workflow support.
ITS	IT Solutions; the managed services provider delivering the Managed AI Program.
LLM	Large Language Model; an AI model trained on large amounts of text that can understand, summarize, generate, and transform language-based content.
M365 Copilot	Microsoft 365 Copilot; Microsoft’s AI assistant experience for Microsoft 365 applications and data, subject to client licensing and configuration.
MCP	Model Context Protocol; a protocol used to connect AI systems to approved tools, systems, or data sources in a standardized way.
MCP Server	Hosted infrastructure that enables MCP-based connections between an AI system and approved external tools or data sources.
MFA	Multi-Factor Authentication; a security method requiring users to verify their identity using more than one factor.
Non-MCP Connection	A supported connection method that does not use MCP but may still allow assisted access to approved applications, workflows, or systems.
Open Office Hours	AI Community sessions hosted by an AI Business Analyst where users can ask AI usage questions, receive prompt guidance, and review examples or tips through live online meetings.
Private LLM	A private, sandboxed large language model environment used for approved workloads that may involve sensitive, lengthy, or high-volume AI use cases.
Professional Services	Separately scoped and billed project work for items outside the standard Managed AI Program stage limits or included monthly service scope.
Prompt	The instruction, question, or context a user provides to an AI system to request a response or action.
Prompt Coaching	Guidance that helps users write clearer prompts, improve AI outputs, and use AI more effectively for approved tasks.
RAG	Retrieval-Augmented Generation; an approach where AI retrieves relevant information from approved knowledge sources before generating an answer.
RAG Server	Hosted infrastructure used to connect AI to approved knowledge sources so users can ask questions against company content, subject to configured permissions and source-system access controls.
Service Desk	The support channel used for workspace access issues, deployed solution failures, connection issues, integration errors, and other operational issues included in support scope.
Shadow AI	Use of AI tools or services that have not been approved, governed, or managed by the client organization.
SSO	Single Sign-On; an authentication method that allows users to sign in using approved organization credentials.
User Prompt Insight Reporting	Stage-based reporting that provides visibility into user prompt activity and AI usage patterns where included and supported by the platform.
vCISO	Virtual Chief Information Security Officer; an advisory role that supports cybersecurity strategy, risk management, and security governance.
Workflow Automation	The use of software to execute repeatable tasks or processes automatically based on triggers, rules, or approved AI-supported actions.

[Return to top]