Prerequisites

Table 1. Prerequisites
Category	Item	Description	Client Responsibility
I. LLM & Tooling Integration	LLM Subscription Details	Large Language Model (LLM) Access: Details regarding the client's LLM subscription, including API keys, endpoints, and any specific model access permissions required for the Agentic AI GenAI to operate. This ensures the AI agent can communicate with the underlying LLM services.	Provide valid LLM API keys, service endpoints, and confirm active subscription status.
	ServiceNow Integration Details	ServiceNow Instance Access: API credentials, instance URL, and any necessary permissions for the Agentic AI GenAI to interact with ServiceNow for task creation, or data retrieval. A service account is required with the respective roles: Call the APIs: SOAP, REST & ITIL roles are required. Publish the KB article in SNOW portal: ITIL role is required. API’s Ticket Details API Closed tickets API Add article API Article details API	Provide API tokens/credentials, instance URL, and ensure appropriate role-based access control (RBAC) permissions are granted for API access.
		SNOW team have to come with some mechanism so that: They with called GenAI Bot connect API to generate the token and use that for Login or Authenticate the user. They have to call GenAI API from service now for updating the existing collection for below actions: for indexing the new generated KB article which is published by SNOW (Add, update or publish) For Removal of indexing of Kb article which is deleted/unpublish from the service now GenAI team provide the API (delete/unpublish/removal)
	Other Tooling (if applicable)	Any other specific tools (e.g., monitoring tools, knowledge bases) that the Agentic AI GenAI needs to integrate with. This includes API documentation, access credentials, and desired interaction patterns.	Identify additional tools, provide their API documentation, and grant necessary access.
II. Workflow & Content Approval	Approved Workflows & Processes	Agentic AI Workflow Approval: Comprehensive, client-approved workflows for all processes the Agentic AI GenAI will automate or assist with. This includes decision trees, escalation paths, and desired actions for various scenarios. Example: ServiceNow ticket creation workflow, Nexthinkalert remediation workflow.	Provide detailed, documented, and officially approved workflows for each use case. This may involve step-by-step descriptions.
	Knowledge Base Access	Access to Client Knowledge Bases: Read permissions to relevant client knowledge bases, FAQs, and documentation that the AI agent can use for self-service or information retrieval.	Grant read-only access to relevant knowledge repositories or provide curated content for ingestion.
	Communication Guidelines	AI Agent Persona & Tone: Guidelines on the desired tone, persona, and communication style for the AI agent (e.g., formal, friendly, empathetic, technical). This ensures consistency with client branding and user experience expectations.	Define and approve the AI agent's communication persona and specific interaction guidelines.
III. Network & Security	Network Whitelisting	Network Access Configuration: List of IP addresses or domain names for the Agentic AI GenAI platform that need to be whitelisted on the client's network firewalls to allow inbound and outbound communication.	Configure network firewalls to allow necessary inbound/outbound traffic from specified IPs/domains.
	Security & Compliance Review	Security and Compliance Approval: Collaboration with the client's security and compliance teams to ensure the Agentic AI GenAI adheres to all internal security policies, data privacy regulations (e.g., GDPR, HIPAA), and industry standards.	Participate in security reviews, provide documentation on compliance requirements, and grant necessary approvals.
IV. Client Collaboration & Support	Designated Contact Person(s)	Primary Points of Contact: Identification of key stakeholders and technical contacts within the client organization responsible for project oversight, technical approvals, data provision, and ongoing support.	Designate and provide full contact details (name, title, email, phone) for all relevant stakeholders and technical contacts.
	Testing & Validation Support	User Acceptance Testing (UAT) Support: Availability of client personnel to participate in testing, provide feedback, and validate the functionality and performance of the Agentic AI GenAI during deployment phases.	Allocate resources for UAT, provide timely feedback, and formally approve UAT results.