Service desk integrations in 2026: connectors, MCP and metered automation across nine tools

The Model Context Protocol is an open-source standard for connecting AI applications to external systems such as data sources, tools and workflows [Model Context Protocol]. On a service desk page, MCP support means an AI agent can reach other systems through a common interface rather than one custom connector each. monday service and Freshservice both mention MCP on the pages we read [monday IT page] [Freshservice]. Ask which systems are reachable through MCP today, which permissions the AI agent gets, and where its actions are logged.

Because a busy month can use them up. SysAid bundles integration runs as Workato credits, 100K on Professional and 300K on Enterprise [SysAid pricing]. monday service Enterprise lists 250K automation and integration actions per month [monday service pricing]. Jira Service Management Standard includes 250 Rovo credits [Atlassian pricing]. A trial rarely reaches these limits, so ask the vendor to estimate a month of your ticket volume in its own units, and what happens when you pass the limit.

• Identity, for access requests and new starters. monday service names Active Directory [monday IT page].
• Chat, where requesters already work. Halo's virtual agent works in Teams and Slack [Halo AI], ServiceNow lists Teams and Slack among its self-service channels [ServiceNow ITSM], and Zendesk's assistant works inside Microsoft 365 apps [Zendesk newsroom].
• On-call and alerting. monday service names PagerDuty [monday IT page]; the Jira Service Management Free plan includes alerts and on-call schedules [Atlassian pricing].
• Engineering tools, if IT and development share incidents. Jira Service Management has native Jira links and deployment gating on Premium [Atlassian pricing]; monday service names Jira and GitHub [monday service page].
• Discovery, if you keep a CMDB. Freshservice describes agentless discovery [Freshservice IT asset management] and ServiceNow describes Discovery across on-premises and cloud [ServiceNow CMDB].

Our integrations criterion looks at documented connectors, API and MCP support, catalogue size where stated, and fit with common IT stacks. The highest scores in the set are Jira Service Management (9) and ServiceNow ITSM (9) [ServiceNow CMDB page], and the lowest is SysAid (6.5). The per-vendor reasons are on each review and in the side-by-side comparison.

1. Which of our systems have native connectors, and which need the API or a third-party tool?
2. Is integration or automation usage metered, and what would a month of our volume use?
3. Which systems can the AI agent reach, through MCP or otherwise, and with which permissions?
4. Where are integration errors shown, and who is alerted?
5. Which integrations are included on our plan, and which need a higher plan?